A Imporved uTracer switch for anode and screen testing at 1 amp

[Never Mind]

As I posted in the past I have been working on a improved switch for the uTracer3+.

I have made several steps forward and well as a few in reverse.

I have made enough progress now to provide a update that I though may be of interest.

The latest version of my improved switch is now working nicely in my lab and will happily switch the full 400V of a uTracer3  at 1.2 amp current  into a short without compliance being turned on or needed. 

It will easily survive repeated full 1.2amp test pulses without compliance and with no risk of damage into basically a shorted load.

More testing is needed but looking very good so far.

There is too much to post here as google groups is a difficult place to post anything complex that will need editing with time.

I hope this does not brake some rule. if so I apologize in advance.

I have uploaded the project to my web site at

http://bluesystems.dyndns.org/

Let me know what you think.

[Gruber, Heimo - DI]

Nice Job! Thank you so much for your efforts. I'm looking forward to the final release.

Heimo Gruber

---

From: utr...@googlegroups.com <utr...@googlegroups.com> on behalf of Never Mind <bobblue...@gmail.com>
Sent: Tuesday, November 26, 2024 2:28:21 AM
To: uTracer <utr...@googlegroups.com>
Subject: A Imporved uTracer switch for anode and screen testing at 1 amp

 

--
You received this message because you are subscribed to the Google Groups "uTracer" group.
To unsubscribe from this group and stop receiving emails from it, send an email to utracer+u...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/utracer/fe43f7f9-cd8b-47fe-8752-00fe40b662a7n%40googlegroups.com.

[Davo]

Wow Bob I’m impressed!

Looks very promising :)

You really know your stuff, I like your testing facility using an Arduino, really clever!

The back side of the board, can it be pulled off? Would be handy to sevice the main board if needed, or can it create instability?

Btw interesting projects on your site! :)

Gotta work now but keep up the good work!

Dave

Op 26 nov 2024 om 05:20 heeft Gruber, Heimo - DI <H.Gr...@riskexperts.at> het volgende geschreven:



To view this discussion visit https://groups.google.com/d/msgid/utracer/VE1PR06MB6847F22E71A49BEBD4F19716EC2F2%40VE1PR06MB6847.eurprd06.prod.outlook.com.

[Martin Manning]

Great, thanks for sharing your work! 

Does your small board replace the entire uT3 switch as outlined in the snip below? 

Is there any change in the current required by the optocoupler? IIRC the ground bounce issue is strongly affected by that. 

IMO 500 mA capability would be adequate for 99% of users.

[Martin Manning]

PS After reading more carefully I see the optocoupler current is much reduced... great! Hoping it is possible to fit this board into my enclosure.

[Never Mind]

Martian said " Does your small board replace the entire uT3 switch as outlined in the snip below "

Note quite.

 

I left off from the module the the storage discharge circuit function of R34, C35, T15, R35.

So R34, C35, T15, R35 and D15 has to stay.

The discharge function took up too much space to add needing a large (for this module) power resistor.

Also the current discharge parts are rated with enough voltage with the MPSA44 easily handling 400V or 500V right to the limit with a base emitter resistor as the uTracer3 currently has (R34 to the PIC I/O port) .

Seem to function fine as designed so I did not see a need to replace this function.

Finally the design intent of the module was the module could be added and ALL current uTracer3 parts left in place on the uTracer3 PCB to allow easy removal of the module in the future should you want to.

When the 4N25 opto coupler is unplugged to plug in the new module this disables the current Utracer3 switch circuits on the main board. (well that is the plan, not test yet!).

This makes adding the module a much lower risk for a uTracer owner. I have no desire to cause any uTracer owner to trash their unit.

There was a struggle for space to make sure the board was narrow and would fit between the two large capacitors. 

I also wanted the PCB to be short enough to stay low and so  it had to fit between the other two high parts at each end, T17 and T14.

The idea was to not be much higher than the large capacitors and so fit into most uTracer enclosures.

This space restriction limited what parts could be used and the high voltage isolation requirements in some parts of the module further constrained the layout.

I have in the back of my mind that the uTracer3+ should be easily extendable to 500V operation with this new module.

That is a future idea and I am not testing or spending much design effort on that right now. Maybe some day.

The module has three current setting options set by jumper pins, 200mA , 600mA and 1A so you can have what you want.

Set it to 200mA for low current uses or set it to 1A and test sweep tubes or rectifiers with ease.

I may make a little current range changing module to allow setting the current sense resistors R45, R20 from the front panel.

As long leads will not work it will use some sort of relay either mechanical or electronic.

That will allow the high current sensitivity and accuracy of a stock uTracer3+ and the 1A ability of the uTracer6 with the better voltage accuracy of a uTracer3+.

Be great for sweep tube, rectifier testing or any other low impedance testing.

However time will limit how much dreaming becomes real. Sigh.

When you look at all that Ronald has got done in his spare time it is pretty impressive.

Bob

[Martin Manning]

Bob wrote: 

I left off from the module the storage discharge circuit function of R34, C35, T15, R35. 

So R34, C35, T15, R35 and D15 has to stay.

So it would be as shown below? Note D15 on the older schematic is D4 on the 3+ schematic. Seems like you would want to break the old path from the reservoir capacitor to the anode/screen terminals, no?

I devised a pin header and jumper arrangement which could be added to the current PCB to select low (the standard 220 mA) and high (500 mA) current range. One would have to open the enclosure to move the jumper, and reset the sensing resistor value in the cal file, and maybe recalibrate, but that would be a workable solution if the best possible resolution were needed. I haven't installed it yet, as the 500 mA configuration seems to be working well for me.

.

[Never Mind]

Martin wrote :"Seems like you would want to break the old path from the reservoir capacitor to the anode/screen terminals, no?"

Removal of a few more part may be required.

It would be nice if all the old parts could be left in place to make removal of the module easy if desired.  Testing will tell.

With the opto coupler removed the control path to turn on the old switch is no longer present so the old parts should just stay off and only "come along for the ride".

However in the world of high voltage switching and unexpected switching transients the old circuits may cause mischief if left in place.

The old T17 and T18 will add some stray capacitance and that may become a issue.

It may come to pass that T17 and T18 must be removed to isolate all the old circuits from the new module's switched output.

You could also cut the track from T17 to the fuse however I dislike suggesting track cutting in general.

[Never Mind]

Martin wrote :"I devised a pin header and jumper arrangement which could be added to the current PCB to select"

I have also been thinking on this. 

For me taking the uTracer out of the case to change ranges seemed like too much trouble to actually happen.

I figured I needed some way from the front panel to switch ranges if I was ever to use a range change function.

One idea is to use a ultra low resistance PMOS transistor in series with one of the range resistors.

The PMOS source goes to ground and the drain to the "bottom" of the lower value sense resistor.

The PMOS would then be controlled by a -10v DC signal from the panel and so no high speed signals would go anywhere.

Something like the attached part that has a on resistance of 0.0098 ohms. They are low cost and small. Some few extra parts are required to protect and control the gate.

Some sort of trim pot for the second range would be provided so you would software cal. on one range and use a trim pot to put the second range into cal. Avoids having to have two cal files and keeping them straight.

The present GUI has no allowances for users having more than one uTracer in terms of cal. file management nor any allowance for two range units.

For my setup with a Utracer3 and a uTracer6 I setup all the uTracerJS software files twice in two separate directories so I can make sure I get the correct cal and setup files for each unit as well as keeping the data files separate.

Seems crude but it works just fine for me.

The same would work with two range units.

[Davo]

Hi Bob!

Any news or progress on HV switch?

Kind regards,

Dave Achterberg

Op 27 nov 2024 om 02:18 heeft Never Mind <bobblue...@gmail.com> het volgende geschreven:



To view this discussion visit https://groups.google.com/d/msgid/utracer/b651e5a0-c615-4680-865c-aeba84317a63n%40googlegroups.com.

<si4425ddy-1.pdf>

[Never Mind]

Dave

It has been a while now alright. 

I did a revision the the PCB in the fall to capture change and the new boards got caught up in a long postage strike, then the xmass rush.

Maybe the PCB will arrive in a week or maybe not. 

Post office uses it's own special calendar so it can be hard to tell when.

I will update when they arrive.

In the mean time I been having a lot of fun matching tubes with Ihor's latest revision of uTracerJS.

You should check it out. The Current version of his quick test is wonderful for matching tubes.

For both triodes and pentodes it can now automatically  find the correct Vg1 for whatever Ia target you ask for then spits out all the important parameters.

Makes matching batches of tubes a joy.

Bob

[Davo]

Hi Bob

No rush! I was just curious.

At the moment I haven’t got much time to focus on my uTracer.

To much things going on at the moment.

I was following the thread about the quick test upgrade, it looks very promising though!

For me it’s the most powerful feature of the uTracer!

Ronald did a good job on it, but you guys took it to the next level.

Especially Ihor for his programming skills, I’m a real noob as it comes to programming haha

Without the input from you guys it probably didn’t evolve to this level…

I’m in some kind of gap at the moment with my uTracer.

I have Nick’s Matrix board & Firmware upgrade which are for me both great addon’s! 

Plus i find the uTmax GUI also really cool! All settings & patching in 1 preset. It tests automatically for shorts & leakage before testing.  Resizing windows, theme’s, exporting models & plots are great also

The onboard heater with the new firmware & mod has improved quite a-lot! No need for an external power supply, but i have to do some good extended testing for that..

What I dislike is the minimalistic quick test…

As far as I know Ronald & Ihor’s GUI’s don’t support the benefit of the Firmware & Matrix board.

So i’m stuck in a gap…

Op 4 jan 2025 om 21:24 heeft Never Mind <bobblue...@gmail.com> het volgende geschreven:



To view this discussion visit https://groups.google.com/d/msgid/utracer/a0be7c2a-5374-4eeb-a784-7baa5b15ebe6n%40googlegroups.com.

[Never Mind]

Dave

In my assessment is the uTmax leads in convenience.

So for a general purpose tester uTmax is quick and easy.

The matrix is fast and easy for more common tubes but not so great for the really weird stuff. Think 117P7 for example.

So for many users the matrix is great but for me a not so much in terms of quick and easy flexibility for the weird stuff in my collection..

The uTracerJS software leads the way as a accurate lab tool in my experience.

As I use my uTracers only as lab tools so connivance is less of a concern than accuracy and flexibility.

First the uTracerJS has improved correction factors that allow more accurate measurements under all conditions.

1) These improved correction factors provide a modest boost in uTracer3+ accuracy that most uses may not be concerned about.

They provide a good boost in accuracy for a high current utracer3.

They are in my experience absolutely required for a utracer6 to produce usefully trustworthy data at higher current levels below 100 volts. 

Without them on a uTracer6 you can experience 50% to 100% errors below 75 volts at high currents. This is in the key "knee" region in high power and current pentode curves.

This is in my experience very important if you are trying to create accurate spice models of high current tubes.

2) The correction factors now in uTracerJS allow the user to make changes to the HV switch circuit and account for those changes in the user settable correction factors.

Without this ability making any changes to the uTracer switch circuit will seriously degrade the uTracer accuracy at higher currents and lower voltages.

So if you want to upgrade your uTracer for higher currents or reliably at higher current testing uTracerJS is required to reach high accuracy.

3) utracerJs quick test now uses a very effective averaging system for measurements that reduces error levels to 1/2 to 1/4 of non averaged results.

This results in better tube matching and more repeatable results.

It now automatically finds the correct Vg1 voltage for your test conditions and then goes on the automatically measure all tube parameters to high accuracy.

I am finding the new quick test far more accurate than what I achieved in the past for in making matched set of tubes.

4) uTracerJS lets you run two uTracers at the same time on the same PC. I run my uTracer3 and my uTracer6 at the same time from the same PC.

Just setup uTracerJS in two separate directories and plug in two serial ports. Set each to different IP "ports", I use port 9300 and 9600 to make it easy to see what machine I am connected to.

Works nicely and is very handy if you have a lot of tubes to test.

In the end no need to be stuck, just install both uTmax and uTracerJS. 

That's what I have done. 

Pick what works for the tests you are doing right now.

I do admit however I have not used uTmax in a long time (from when I got a uTracer6 amost 1 year ago) as the greater accuracy of uTracerJS is just too compelling on my utracer3+ and really absolutely needed with a Utracer6 in my experience.

Bob

[Davo]

Hi Bob

Do you have a final parts list for the HV Switch?

I have to order some stuff at Digikey, so it’s more costs effective to order extra parts

If there are parts for which you have different options then I will order those as well

Dave

To view this discussion visit https://groups.google.com/d/msgid/utracer/d5b6147c-1c11-456a-8452-7bce73b0a013n%40googlegroups.com.

[D.A.R Achterberg]

Hi Bob,

I looked at the schematic on your site and took the components the were listed on the schematic…

Are there any adjustments or parts that I maybe have to order extra just to be sure..

Those jumpers aren’t available at digikey as well for Q1

R1= 3.74K 1% (1206)

R2= 18.2R 1% (1206)

R3= 1.37R 1% (2010)

R4= 1.37R 1% (2010)

R6= 47R 5% (2512)

R7= 100R 1% (1206)

R8= 10R 1% (1206)

R9= 1K 1% (1206)

R10= 18.2R 1% (1206)

R11= 2.7R 1% 1/2W (1206)

C1= 2.2uF 63v Metal polyester 

C2= 47uF 35v  5x11mm

C3= 100nF 100v X7R 0805

C4= 100nF 100v X7R 0805

D1=  US1M, 1000v 1A (DO214AC)

Q1= STP13NK50Z TO220  Discontinued

Q2= STD4NK60Z TO251

Q3= MMBT4401L (SOT23-3)

Q4= MMBT4403L (SOT23-3)

Q5= MMBT4401L (SOT23-3)

ISO1= TLP2745

U1‎ =  IRS44273 (SOT23-5)

3x JP1= BBS-132-T-B  not available 

2x JP2= BBS-132-T-B  not available 

Op 15 jan 2025 om 12:13 heeft Davo <djda...@gmail.com> het volgende geschreven:

Hi Bob

To view this discussion visit https://groups.google.com/d/msgid/utracer/8D1B8D52-5AB0-4001-97B1-413A093C96E8%40gmail.com.

[Never Mind]

Dave

Thanks for your interest.

I have to caution you that the design is still  in development and must be confirmed on a new PCB.

The schematic on the WEB site is for the first revision and had a number of serious issues.

As I said on the WED site...

"The first version of the improved uTracer3+ switch proved to be a excellent power oscillator for a few seconds time."

Then it went bang! 

In design if it is all going smoothly you probably are not doing anything new.

So do not use this first revision.

You will note a spice simulation schematic at the end of the WEB post for the second revision of the improved switch.

The second version is very different from the first version.

Although I have proto typed this second revision on the bench it is still needs to be fully tested on the new PCB.

So I strongly suggest not using any of my design until I can confirm that it works correctly.

Bob J.

[Never Mind]

Finally the post office came through with my second revision PCB for the improved uTracer3 switch.

Today I built a sample of this second revision PCB of a improved utracer3 switch to test and debug on the bench.

The good news it has passed all tests I have put it through.

My test setup used a simple Arduino blue pill circuit to generate the uTracer3 pulses to simulate the operation of a real uTracer3.

This allowed the switch design to be evaluated on the bench without putting my uTracer3 at risk.

I used a 100uF 500V capacitor charged by a 5K 10 watt resistor from a high voltage bench power supply to simulate  the utracer3 capacitor charging circuits.

The 100uF 500V capacitor was selected as it is the same size as used in the real uTracer3 storage capacitor.

This provided a realistic mimic of the utracer3 power supply characteristics.

Finally the switch was terminated into a large bank of resistors that worked out to 4 ohms with 50uH paralleled with a 1k resistor placed in series..

The 4 ohm resistance is close to the value of the current sensing resistor of 4.7 ohms used in a uTracer6 and so the switch should be seeing a total short circuit impedance close to a real utracer3 setup.

The test results.

The switch design allows for 3 current limit settings, each set about 20% above the desired uTracer3 test current.

1) 200mA, standard uTracer3 test current

2) 600mA, high current uTracer3 test current

3) 1A, standard uTracer6 test current

I started at the 1 amp setting so the current limit should be about 1.2A using a low 15V power supply voltage. Safety first!

The attached picture shows the result as a nice clean pulse that stops at the desired current limit of 1.2A. 

The droop you see in test current on the right side is due to the utracer3 storage capacitor discharging during the 1mS test pulse.

The supply was then increased to 21V. Current remains constant at 1.2 amps and you now see the droop is almost gone on the right side as the storage capacitor now has almost enough energy for a full 1.25mS pulse time.

Next test was 450V at the 200mA setting. I used 450V to test with to insure there was some margin in the design over the utracer3 400V limit.

The pulse looks nice and clean.

On the test at 450V @ 600mA. The pulse still looks nice and clean at a steady 20% over 600mA current limit.

Finally testing at 450V @ 1A. This was the scary bit. LOL.

The pulse still looks good holding the current to exactly 1.2A or 20% over 1A test current current limit.

At the 450V level there is now some over shoot of the current limit at the leading edge developing.

Measurements show the over shoot is under 1uS wide and reaches a maximum current level of about 6 amps.

The data sheet safe area for the MOSFET I used states a 12amp limit for a pulse under 10uS with 400V applied so we should be fine at this level.

Remember we are testing into a full short circuit at 450 volts so this is pretty hard on the switch and not really what daily use will look like.

A stock 200mA uTracer3 switch will not likely survive a short circuit with compliance turned off even below 100V.

Keep in mind this switch design is repeatedly surviving in my testing at a full 1.2A current short circuit without compliance turned on at over the maximum test voltage a utracer3 can supply.

This seems like a worthwhile improvement over the stock utracer3 switch and should allow a uTracer3 to test at 1 amp current levels with safety.

As a bonus the new switch design will reduce the grid voltage disturbance issue by a factor of 10 by drawing only 1/10 the current from the PIC I/O port.

Next I will build two more boards and install them in my utracer3 to see how it goes in real use.

If testing in a real uTracer3 goes well I will post the circuit diagram, BOM and the PCB Gerber files for anyone interested.

[Walter Trovò]

Great job indeed!

Waiting for the result of the extensive test!

BTW, you could consider making your PCB available for purchase on your PCB maker, I guess it could significantly lower the price…

Walter

On Jan 21, 2025, at 6:15, Never Mind <bobblue...@gmail.com> wrote:



To view this discussion visit https://groups.google.com/d/msgid/utracer/0a56f544-1005-4bb5-85df-73c56cb3ce2dn%40googlegroups.com.

<uTracer3 switch top.jpg>

<15V 1A 4 ohm load.jpg>

<450V 200mA 4 ohm load.jpg>

<uTracer3 switch bottom.jpg>

<21V 1A 4 ohm load.jpg>

<uTracer3 switch rev10 rat nest.jpg>

<450V 600mA 4 ohm load.jpg>

<450V 1A 4 ohm load.jpg>

<3228AB10-PCB.jpg>

[Davo]

Hey Bob!

Congrats!

It looks very promising, titanium works! I really admire your skills to make it all happen.

Can’t wait to see the results in real life on the uTracer

Keep up the good work!

Dave

Op 20 jan 2025 om 22:15 heeft Never Mind <bobblue...@gmail.com> het volgende geschreven:



[Never Mind]

Today I integrated the new switch board into my utracer3 anode circuit to see what happens in "real" hardware.

Happily it fit right in to the 4N25 socket as designed. That was a relief, I was concerned about how tight on space the design was.

I have so far only installed it into the anode circuit until I can fully debug the design.

One nice feature is nothing in the original uT3 has to be disturbed to install the board so for testing the uTracer3 can be restored back to original simply by unplugging the board and putting back the 4N25 into it's socket.

This has turned out to be a very helpful feature in testing.

On to testing. 

I did a 2V to 10V sweep and found missing samples. What have I missed, it worked perfectly with the arduino simulator?

Next I tried a 2V to 100V sweep seeing below 10V and above 50V missing samples.

As there were no fireworks I suspected things were "mostly" working as designed. What had I missed?

My design uses a modern high speed logic opto coupled flying switch powered by a flying capacitor that is charged between samples. 

This is similar to how uT3 design except the uT3 is based on a old school 4N25 opto.

The modern switch I used requires about 3mA current at all times. Ronald's design only needs power when the switch is active.

So the main difference between designs is both how the capacitor is charged and how much current the switch driver uses.

I used the modern part as it is very fast, requires low current to drive the LED and provide a fast high current output designed to drive a MOSFET.

In Ronald's uT3 design the flying capacitor is powered by a circuit that bleeds power from the HV main storage capacitor. This is a novel way to do things and is very simple.

In my design I charge the flying capacitor from the +15V supply when the switch is off and the the anode is floating. 

This is done by a pull down MOSFET connected to the switch output that grounds through a 47 ohm resistor the anode connection. This is the conventional way to charge a flying capacitor in a switch mode power supply so I make no clams to originality. I note this is also how the uT6 now does this function.

I based my design for charging the flying capacitor on a some of Ronald's WEB site notes where he says "charged to 10V by R7 during the 10ms “charging pulse” that precedes the measurement pulse"

I took this to mean that the charge pulse would be a single 10mS pulse right before the measurement pulse. This is how conventional flying switch designs normally work.

The capacitor I used was designed to power the switch for about 10mS time. Lots of time I thought for a 1.25mS pulse with some mS lost before the measurement pulse starts.

What I found was the the uT3 to charge the capacitor  applied a pulse train at a 300uS pulse rate for about 28mS. The duty cycle was about 93% on.

To make things more complex I found that the charging pulse was as much as 44mS before the measurement pulse allowing time for my flying capacitor to discharge.

This was the cause of the missing data.

It seems I made too many assumptions from Ronald's WEB site notes and should have spent extra time measuring uT3 wave-forms at the project start. Oh well I do tend to be in a rush to get started.

I increased the capacitor I used from 2.2uF to 24uF and got a very nice result in a 2 to 100 volt sweep. 

Ok so we are making progress.

Next I tried a 250V to 400V sweep figuring "we got this".

No such luck. Now I see missing data above 350 volts?

I look at the flying capacitor charge pulses and find to my dismay that above about 350V the charge pulses have moved away from the measurement pulse by 126mS.

In fact the charge pulses are now AFTER the last measurement pulse NOT before the present pulse at all. This results in a 163mS delay from when the flying capacitor is charged and when a measurement is taken.

This is a design in the firmware I had not expected and a simple increase of flying capacitor is not going to correct this issue.

As the uT3 PIC firmware is so variable in timing for charging, a conventional design flying switch circuit will not work driven by this firmware.

I am going to have to integrate flying capacitor charging logic into my switch design and no longer depend on PIC firmware to do this task.

This will have the advantage of making charging the switch's power source time consistent across all measurement voltage and currents and may reduce a possible source of measurement variation.

We will see in a few days when I get the required changes made to my control circuit.

[Never Mind]

Walter said "BTW, you could consider making your PCB available for purchase on your PCB maker, I guess it could significantly lower the price…"

I had not thought about that but would be happy to do so. I guess it will make it easier for some to order.

Not sure any cost reduction you get will make your day as I only paid 4USD for 5 pieces of this very little PCB so it will not break the bank.

I think they make these little PCB from left over scraps as I normally pay between 2USD and 4USD for 5 pieces when the boards are very small.

I can not imagine how they are so cost efficient as the quality I have found completely acceptable.

All the money is in the shipping. LOL.

We are not there yet with this project but seem to be getting closer.

Bob J.

[Walter Trovò]

Thank you for the update, Bob.

What about using a tiny isolated DC-DC module?

Most of them are based on a very simple, unregulated Royer topology.

I opened one from Wurth and I clearly recognized the push-pull driving stage, a toroidal coupled inductor and two rectifiers.

No resonant capacitor at primary, I guess it works by core saturation with two BJTs.

Isolation is typ. 1kV, and they are very cheap…

Walter

On Jan 25, 2025, at 9:13, Never Mind <bobblue...@gmail.com> wrote:



To view this discussion visit https://groups.google.com/d/msgid/utracer/ef58cf38-d45b-4a1c-b619-43d4a37c6334n%40googlegroups.com.

[Davo]

Looks very promising Bob!

And yet again a very nice detailed reportx 

i’m really looking out for following results

Dave

Op 25 jan 2025 om 01:13 heeft Never Mind <bobblue...@gmail.com> het volgende geschreven:



To view this discussion visit https://groups.google.com/d/msgid/utracer/ef58cf38-d45b-4a1c-b619-43d4a37c6334n%40googlegroups.com.

[Never Mind]

Walter said " What about using a tiny isolated DC-DC module?"

Good suggestion and I have used this on past designs using a flying switch that needed to be suspended for a long time (long in switch mode land is 10mS or more)

A few concerns with going this way.

1) Noise. Ronald shuts down everything in the uTracer to make a measurement to make things as quiet as possible and this is a smart move.

I would have no way to shut down the DC to Dc converter unless I added some more "goo" to do this in hardware.

2) Space. I wanted the PCB design to fit right into a uT3 with out "hacking" the unit to pieces. I am already up against the wall space wise on my little board.

I am close to getting there now but this could be seen as a fall back.

In a perfect world I would have the PIC source code and the fix would be quick and easy. I however must play the hand I have been dealt and try to make the current PIC code work.

[Never Mind]

Today's update is progress if not yet perfection .

I changed the module's logic from charging the flying capacitor when the PIC sends the discharge signal (how the stock uT3 works) to charging the flying capacitor whenever there is no measurement pulse.

This means the flying capacitor is always charged and ready for a measurement pulse.

The result is it no longer matters when in the sequence of various the hardware signals the measurement pulse is applied.

The result was perfect operation for 2 volts to 400 volts and up to 1 amp current. 

I attached screen shot of a sweep of a 333 ohm resistor. It looks pretty good.

I also attached a screen shot and the  UTD file for a sweep from 2V to 120V into a 100 ohm resistor.

With it is a screen shot of the error plot and the XLS sheet showing error levels in a 2V to 100V sweep into a 100 ohm load up to 1amp.

This sweep was taken with no extra averaging. The peak error for one sample was 5% with a majority of errors under 1%.

This is before I spent much time in calibration so maybe it can get a bit better. This level of error already eats my uT6 for lunch in term of low voltage accuracy.

Whats the catch?

In order to deal with the PIC firmware not consistently providing a charge pulse BEFORE a measurement pulse I moved the timing of flying capacitor charging to the hardware's control.

As the hardware is very simple this meant a very simple logic decision.

If NOT a measurement pulse THEN charge capacitor.

It is not quite this simple as I had to insure that the HV switch was always fully off before the capacitor circuit is enabled or shoot through currents would occur in HV switch and the MOSFET that charges the flying capacitor.

This I did with some simple changes to the capacitor charging MOSFET's gate drive signals. Spice is your friend when doing this sort of stuff.

What is the catch?

Since the flying capacitor is charged by grounding the anode or screen switch output this means if the anode or screen is connected to the cathode line that sits at +19V there will be a lot of current when the capacitor charge command is received.

This is how the uT6 works and can in fact blow the uT6 anode or screen fuse if you short the anode or screen lead to the cathode and then make a measurement.

This because as PIC tries to change the flying capacitor in a uT6 by grounding the anode or screen leads a lot of current flows from the cathode through the anode to cathode short.

I do not like this much from a engineering point of view but in practice it is not a issue that has ever bothered me.

Now in today's module design with the hardware is controlling the capacitor's charge MOSFET, turning it on whenever there is NOT a measurement pulse.

The result is any time a anode or screen lead is shorted to the cathode the anode or screen fuse will blow even with the uT3 sitting doing nothing.

Thus we come to the catch and not a great feature of the present version of my switch.

I am thinking on how to overcome this issue.

A settable PTC fuse in the MOSFET charge circuit that will open before any fuse blows is today's preferred fix.

We shall see. Sure wish I had access to that PIC code.....

I attached today's schematic for anyone that wants to see where we are BUT remember this is not a finished design so use at your own risk.

[Walter Trovò]

Oh, brilliant! Great job!

Walter

On Jan 27, 2025, at 5:29, Never Mind <bobblue...@gmail.com> wrote:



To view this discussion visit https://groups.google.com/d/msgid/utracer/0552f247-fd6f-4e8c-92a2-15a358a60657n%40googlegroups.com.

<3228AB20.jpg>

<2-400v 333 OHM 1A.png>

<test sw-int dis Va=2-120 1A 100OHM.png>

<100 ohm 1amp uT3 anode.jpg>

<2-125v_100 ohm 100p anode.xls>

<test sw-int dis Va=2-120 1A 100OHM.utd>

[Never Mind]

I updated the PCB design to change the capacitor charging logic and added a PTC to the capacitor charge circuit to limit the over current event in the case the anode or screen lead is shorted to the cathode.

I was quoted 2USD for quantity 5 boards plus shipping from jlcpcb.com. I find them as a vendor low cost, fast and and the quality is fine in my experience.

The warnings.

1) This is a small, crowded, double sided surface mount board with a few challenging parts to solder like a 5 pin SOT-23.

Sorry but this was required to make the board small enough to fit into the space available in a uT3

There are however no super hard to solder parts like ball grid arrays so it can be build with noting but a fine tip soldering iron and fine wire solder.

Before you consider building a copy ask if your soldering skills (and eyes) are up to the task.

2) We are still at the bleeding edge. This finial PCB version with all the fixes has not been built by myself yet.

I have modified my current 3228AB10 version PCB for the changes above and it works well in the lab so I believe the 3228AB20 version will be 100% clean, however.....

So I was reluctant to post all the design files as I can not personally confirm the PCB is 100% clean.

If anyone out there wants to live on the bleeding edge and build one before I get around to ordering another set of PCBs let me know and I will post all the design files for you.

3) Unlike Ronald's stuff this module has not been subject to years of testing and sales. There may be wrinkles.

4) It will make testing resistors below 50 ohms impossible as the module will assume there is a short between the anode or screen lead and the cathode and so prevent measuring pulses.

This is nice in that the module will not allow a measurement pulse if it sees a anode or screen to cathode resistive short in the setup from wiring errors or shorted tubes.

This should not effect testing tubes at all at any impedance as current can not flow from a positive cathode to a negative anode or screen.

Testing any tube impedance will be possible without difficulty as they only pass current in one direction only.

The good part.

1) The modules is designed to withstand a full 400V measurement pulse into any impedance without the need for compliance to be turned on for protection.

This allows the full 1A test current to be delivered at any anode or screen voltage into a test impedance. 

No more problems with compliance errors testing near the maximum test current.

There should be no danger of destroying your uT3 by turning off compliance to get those higher current readings.

2) With 2 jumpers the module can be quickly set to current limit at 3 levels of test current for flexible use.

200mA, stock uT3

600mA , high current uT3

1A , uT6

3) The module is designed to be installed so that it can be disconnected easily and removed with little effort.

There are no PCB track cuts needed. There are no parts that need be removed from the uT3 main board.

Normal uT3 operation can be quickly restored by simply unplugging the module and inserting the 4N25 back into the socket.

This is handy during testing of the module if there are troubles.

4) Installation of the modules requires the 4N25 be pulled from it's socket and the module plunged in, then 4 wires to be soldered to the uT3 main board.

Two diodes and two 1/4w resistors are added on back of the the main board to suppress voltage flyback in case of a short or loose connection during testing. 

These are a good addition for any uT3 as the uT3 can be vulnerable to high voltage fly-back from the RFI suppression inductors if there are loose connections during testing. 

This is especially true in the case of the high current uT3.

5) The module gives you the better current capability and accuracy than a uT6.

The uT3 has close to 4 times the accuracy in Va and Vs readings as a uT6 and this really helps accuracy when testing below 100 volts.

Because the uT3 uses a 100uF storage capacitor and the uT6 only 50uF total there is 1/2 the voltage droop during a high current measurement in a uT3. The lower voltage droop improves accuracy and increases maximum test current at lower voltages.

The switch losses in the MOSFET I used on the module are 0.55 ohm VS the 8.5 ohms of switch losses on a uT6 switch. This greatly increases accuracy at higher currents below 100 volts from a stock uT6.

6) With this module power rectifiers (5AR4, 5U4, etc) can be characterized with accuracy at the high levels of peak current they normally operate at.

This makes confirming rectifier life left and current capability easy while answering the question "is my power supply voltage low due to a  rectifier issue".

The module also allows high power sweep tubes to be measured at the high currents and low anode voltages they can deliver.

Let me know if anyone wants the design info to build boards before I have time to order and build a finial version myself.

[Walter Trovò]

Hi Bob,

I am in! Is the JLCPCB code 3228AB20?
I have an order to place with them, I will add a couple of your boards.
SMT devices are no problem to me, used to work with 0402 or even 01005…
The problem is that my uT3 is in another Country, and I will be able to test it in a few months…
Are you going to share BOM and final schematic?

Thank you,
Walter



> On Jan 29, 2025, at 1:29, Never Mind <bobblue...@gmail.com> wrote:
>
> 3228AB20

[Never Mind]

Walter

I have attached all the design files in a single ZIP file " 3228AB20-IP.ZIP".

If you simply upload this ZIP file to JLCPCB's WEB site it will automatically create all the information needed to order boards. 

Considering this board was not created with the beginner builder in mind let me know if you run into any trouble or see errors.

I freely admit I am not as detail orientated as Ronald so "caveat emptor" applies.

BTW

The two MOSFET are critical from a timing and ruggedness perspective so I do not recommend any substitution of these two devices.

Also the fT, hFE and ruggedness of the Onsemi  MMBT4401WT1G is important so again I do not recommend substitution of this part

The TLP2745 and IRS44273 are single source as far as I know.

The header for the module to uT3 main board opto socket connection needs to be a exact size that the Samtec BBS-132-T-B provides for.

The resistors and capacitors are mostly jelly bean parts that can come from various vendors.

Watch the 2.7 and 1.37 ohm resistors as they are designed for impulse applications.

I purchased all parts from stock at mouser and expect Digikey may well have them too.

I will at a later date post pictures on adding the module to a uT3 main board.

[Never Mind]

I should have noted the schematic diagram is file "3228AB20-SCH.PDF" and the BOM file "3228AB20.BOM"

[Walter Trovò]

Thank you Bob!

I will submit the Gerber files to JLCPCB and I will get the parts!

--

You received this message because you are subscribed to the Google Groups "uTracer" group.
To unsubscribe from this group and stop receiving emails from it, send an email to utracer+u...@googlegroups.com.

To view this discussion visit https://groups.google.com/d/msgid/utracer/8efe8628-d575-429f-ad19-639db46774e5n%40googlegroups.com.

[Deejay Davo]

Hi Bob!

I have ordered the pcb's. I'm busy placing an order at Digikey, but I'm not sure about SMD1812B035TF/60, they have this one available: SMD1812B050TF/30 

is this one okay?

And wire pad 60 & 80 is that just wire?

Op woensdag 29 januari 2025 om 13:52:30 UTC+1 schreef bobblue...@gmail.com:

[Davo]

Btw I have 2 gmail accounts, , so it can be confusing sometimes haha

d.a.r.ac...@gmail.com & djda...@gmail.com are the same

Regards Dave

Op 1 feb 2025 om 22:43 heeft Deejay Davo <djda...@gmail.com> het volgende geschreven:



--

You received this message because you are subscribed to the Google Groups "uTracer" group.
To unsubscribe from this group and stop receiving emails from it, send an email to utracer+u...@googlegroups.com.

To view this discussion visit https://groups.google.com/d/msgid/utracer/eb7b573c-5268-4ece-aef4-423dd8533c5cn%40googlegroups.com.

[D.A.R Achterberg]

Hi Bob,

Are there some parts I need to order extra?

I don’t know if the design is rugged?

And since the design is fairly new…

Or are there some other parts I need to order that you have in mind for an upgrade?

Op 1 feb 2025 om 23:40 heeft Davo <djda...@gmail.com> het volgende geschreven:



To view this discussion visit https://groups.google.com/d/msgid/utracer/5D7EA065-A1AA-46DC-9BFD-394BF25FF86E%40gmail.com.

[Never Mind]

Note sure who you are ordering from...

SMD1812B050TF/30 is too high a trip current.

Mouser and Digikey both have SMD1812B035TF/30 that is fine.

Yes the wire pads are a PCB feature that wires solder to.

I have to enter the wire pads into the schematic to make them appear on the PCB. 

The result is they also appear on the BOM.

So nothing to buy.

 https://www.mouser.ca/ProductDetail/YAGEO/SMD1812B035TF-30?qs=0lSvoLzn4L8pXXBFdnPTwA%3D%3D

[Never Mind]

The design is very new. It works fine in my lab but it is early days.

If you want to be one of the first to build it you may find some wrinkles I have not ironed out yet.

If you are unsure about dealing with uncertainty in a very new project you may be best to wait for a few others to build before jumping in your self.

Your call.

[Never Mind]

Dave

Good to know and yes you did manage to confuse me. (Not hard to do i admit)

Bob

[Big Josh]

Any good PCB design software, including KiCad, has the option for "exclude from bill of materials" on any item you add to your schematic. In KiCad you just double click whatever component you want to change those details for.

--

You received this message because you are subscribed to the Google Groups "uTracer" group.
To unsubscribe from this group and stop receiving emails from it, send an email to utracer+u...@googlegroups.com.

To view this discussion visit https://groups.google.com/d/msgid/utracer/5a320fb5-163d-4997-a381-111e2a159bb3n%40googlegroups.com.

[Davo]

Thanks for the info Bob!

The order on digikey has heen placed…

Op 4 feb 2025 om 02:28 heeft Big Josh <shif...@gmail.com> het volgende geschreven:



To view this discussion visit https://groups.google.com/d/msgid/utracer/CABi0_%2BPaULQ-TZfqmijk9hcrZgsx8xmkFXuA9ZCDFKcPJUJCAQ%40mail.gmail.com.

[Walter Trovò]

PCBs arrived…

Walter

On Feb 5, 2025, at 3:15, Davo <djda...@gmail.com> wrote:

Thanks for the info Bob!

To view this discussion visit https://groups.google.com/d/msgid/utracer/9FBD8F11-1BAD-4CEA-B4A1-F70C5C0F2F1D%40gmail.com.

[Never Mind]

Walter

Looks good.

I am still waiting on postal delivery of my sample boards.

In the zip file there is a PDF file of the PCB layout that can be useful to answer silkscreen questions for any bit that do not print 100%.

[Davo]

I have received the boards & populated them, I have have to wait to till my board is functioning properly…

Had quite some issues because I burned my uTracer due some stupid mistakes lol

But I really like your stile of engineering Bob!

Kind regards,

Dave

 

Op 9 feb 2025 om 17:23 heeft Never Mind <bobblue...@gmail.com> het volgende geschreven:



To view this discussion visit https://groups.google.com/d/msgid/utracer/3bc1e913-7f5d-430b-9d1e-0a93a9328156n%40googlegroups.com.

[Never Mind]

Nice and tidy soldering. Puts my soldering workmanship to shame.

I tested my modules on the bench before installing. 

This is very important if I assemble them as I will make mistakes! Last module I built I stuffed the dual diodes for the MMBT4401 and MMDT4401s for the dual diodes. LOL

I used for testinh

A fixed voltage 5V and 15V at low current  bench supply.

A variable voltage bench supply of 10-12 volts at 1.5 amp or more current.

A small power resistor of about 4 ohms or so.

A small low current push switch.

A scope that can trigger on a single event (some type of storage scope)

A volt meter

Make sure all power supply are turned off at this time

Connect the negative lead of the 15 volt supply to the ground (GND) wire on the module.

The plus 15 volts to the (P15V) lead on the module. Watch the polarity or you will hurt your module.

Connect the negative lead of the 5V supply to pin2 of the OPTO socket.

Connect the negative lead of the 5V and the negative lead of the 15 volt supply together.

The positive lead of the 5V supply will connect  through a push switch to Pin #1 of the opto socket.

Connect one end of the 4 ohm resistor to the switch output lead on the module marked (TO FUSE).

Connect the other 4 ohm resistor lead to the negative lead of the 0-10V 1.5A supply.

Connect the positive lead of the 0-10V supply to the module wire marked (STO CAP)

Do NOT connect the negative leads of the 0-10V power supply and the 15 volt power supply together or weird things may happen.

Connect the meter across the 4 ohm resistor and see zero volts.

Advance the 0-10V supply in voltage and watch the 4 ohm resistor voltage. It should stay at zero volts (no current).

Reduce the 10V supply to zero.

Turn on the 15V supply.

Measure the voltage with the meter between the switch output lead on the module marked (TO FUSE) and the negative lead of the 15 volt supply

It should be very close to zero volts. This shows the flying power supply charge circuit is working.

Move the meter to the 4 ohm resistor.

Advance the 0-10V supply in voltage and watch the 4 ohm voltage. It should stay at zero volts (no current).

Reduce the 10V supply to zero.

Make sure the two jumper pins are removed from the module.

Connect the scope probe across the 4 ohm resistor, scope ground to the 0-10V power supply ground resistor end. Set to normal mode, trigger on positive pulse edge at about 0.5V

Push the switch repeatedly to connect the 5V supply to OPTO pin#1 to create a short pulse across the 4 ohm resistor. 

The module will only produce a pulse for 10s of milliseconds and then the flying power supply on the module will discharge so you have to capture the pulse on the scope.

Slowly advance the 0-10V supply in voltage and watch for a voltage across the 4 ohm on the scope.

The pulse voltage across the resistor should rise to about 1 volt (about 250mA current limit) and then stay at a steady voltage level as you increase the 10 volt supply.

Reduce the 10V supply to zero.

Add a jumper pin across pin #1 and #2

Pushing the switch repeatedly slowly advance the 0-10V supply in voltage and watch the 4 ohm voltage. It should rise to about 3 volt (about 750mA current limit) and then stay at a steady voltage as you increase the 10 volt supply.

Use a finger test on the mosfet to make sure it is not getting hot. You do not want to cook the MOSFET.

Add the second jumper pin. across pin #3 and #4 so now there are two jumpers.

Pushing the switch repeatedly slowly advance the 0-10V supply in voltage and watch the 4 ohm voltage. It should rise to about 4.8 volt (about 1.2A current limit) and then stay at a steady voltage as you increase the 10 volt supply.

Use a finger test on the mosfet to make sure it is not getting hot. You do not want to cook the MOSFET.

If you get this far you should have a good module.

Let me know if you need a circuit diagram of this test setup and I will draw up something.

Take care.

Bob

BYW

My boards came today. Three cheers for the postal system.

[Davo]

Hi Bob!

Thanks for the compliment, if I had some better quality solder/flux it would be almost perfect…

SMD Soldering is quite easy, for me the trick is putting a tiny bit on solder on 1 pad, then with a tweezer placing the component at the right spot.

Then I place my thumbnail on the component, give it a tiny push downwards & tipping the leg again, so it’s flush with the board, and then I solder the remaining legs.

With IC’s en other parts with more legs I use the same technique again.

Hopefully this can be in use for someone else…

But anyway, all the workmanship credits goes to you Bob!

The design looks very professional,  I wish I had your engineering skills though

  

 Hopefully I can make some time to do a test tonight…

I have lot’s of things going on at the moment.

I’ll keep you updated!

 

Verstuurd vanaf mijn iPhone

Op 16 feb 2025 om 02:10 heeft Never Mind <bobblue...@gmail.com>

To view this discussion visit https://groups.google.com/d/msgid/utracer/ac0b9254-9fb2-4210-b89f-7464213aef5en%40googlegroups.com.

[Walter Trovò]

Work in progress here too!

I populated all 5 PCBs at 70%.

With next Digi-Key order I will complete them all.

I didn’t pay enough attention to all letters in the PN, therefore I bought BAV99 in SOT-323 (ouch!). 

Not a problem actually, I managed to mount them too!

Walter

On Feb 17, 2025, at 20:55, Davo <djda...@gmail.com> wrote:



To view this discussion visit https://groups.google.com/d/msgid/utracer/1F91FB38-36D6-4F8F-85CC-82E36AD78CD9%40gmail.com.

[Davo]

Hi Bob,

All the wiring has been done correctly, but I think I made a fault and burned something?

I accidentally skipped the first step by turning on the 15v supply first, and measured 15v from fuse to ground

After that I turned off the 15v, and I advanced & reduced the 10v supply and the voltage over the resistor stays 0v

After that I tried step 2 with the 15v, and measured again 15v from Fuse to 15v ground

So I aborted the test… 

Kind regards,

Dave Achterberg 

Op 17 feb 2025 om 14:11 heeft Walter Trovò <walter...@gmail.com> het volgende geschreven:



To view this discussion visit https://groups.google.com/d/msgid/utracer/468C7E5F-E545-420D-9D32-3A4AC4F6B8FC%40gmail.com.

[Never Mind]

I attached a block diagram with a simpler series of tests for your module.

If you can run these tests and let me know the results we should be able to find the trouble together.

Just advance in testing until you get a test failure and let me know the results or your questions.

That way we should be able to get to the bottom of any issues slowly but surely.

Bob J.

[Never Mind]

I forgot a detail on my test diagram. 

The P5V and P15V supply grounds need to be common and connected together.

[Walter Trovò]

Hi everyone,

Any progress from your side?

I just completed the 5 PCBs... As said, no way to test them on my uT3+, that is 10000 Km away...

I will soon test them as per Bob's instructions.

BTW, any uT3+ user nearby Tokyo?

Walter

[Walter Trovò]

Hi, 

I quickly tested one PCB and I can see the pulse across the 4R test resistor.

 Actually I didn't have 4R on the spot, therefore I used 5R, and the pulse height is 1.25V, that sounds good (250mA * 5R).

Only the pulse duration seems much shorter than expected: 7 ms instead of 20 -100 ms as per test instructions...

Bob, is the behavior correct?

Thank you,

Walter

[Walter Trovò]

Ah, I tested all 5 PCBs and they behave exactly the same.

Changing the jumpers configuration, I see 1.25V, 3.95V, 6.25V pulse amplitude, that correspond to 0.25A, 0.79A, 1.25A current limits.

The duration of the pulse is always the same.

Maybe 2u2 for C2 are not enough to drive Q1's gate when Q3 is pulling down?

Walter

[Never Mind]

Walter

Sorry Walter I have been distracted finishing a project that I want off my bench.

Congratulation your modules look good.

7mS is fine as in a utracer3 the pulse event is only 1.25mS long.

I tested my modules with a uTracer3 simulator I made with a arduino board to create uTracer control pulses. 

Then I created simplified tests to allow others to more easy do a module check out.

20~100mS is how long the flying power supply will last powering the module when NOT seeing a short circuit.

When tested into a short circuit the module's flying power supply drain is at it's maximum due to the current limiter function of Q3 pulling down the gate of Q1 through R11.

I forgot to take that extra short circuit drain into account when I suggested the flying power supply will last 20~100mS. 

Sorry to lead you astray on the pulse width timing. 7mS is fine into a short circuit.

In a real uTracer3 the flying power supply is charged whenever there is not a pulse event and then the pulse event only lasts 1.25mS.

So the flying power supply will have no problem lasting 1.25mS even in the worst case event of a dead short as you are testing the module.

So your modules look perfect.

By the way I am working on simple current range changing circuit for the uTracer3 so both the 200mA and 1A range can be selected from the front panel.

It will not use relays due to the unstable contact resistance a relay can exhibit.

With dual current ranges the superior noise floor of the 200mA range can be maintained for low current testing while still allowing for a 1A range.

As part of the design there will be a hardware method of allowing for a calibration between the two ranges.

The intent is to calibrate current on one range and then calibrate the tolerance delta on the second range in hardware.

This seems a unfortunate but needed added complexity as none of the software programs allow for the concept of a dual range utracer.

Maybe some day? 

My hope is to see some day both dual current range and dual voltage utracer software support.

This will allow far better resolution below 100V on uTracers for testing rectifier diodes and sweep tubes at high currents and low voltages.

In the mean time with uTracerJS by installing several instances of the program allows for support of several test current level configurations of a utracer. 

I bit clunky of a software workaround but works nicely in practice.

Unfortunately there is still no way to support a dual voltage utracer without post processing the output results to allow for a range "correction".

The post processing requirement breaks the use of quick test tools and other functions so is not really that useful in practice. 

This is pity as low voltage accuracy is the achilles heal of the uTarcer platform when measuring low impedance devices like rectifiers and sweep tubes.

Bob

[Never Mind]

Walter

BTY do not omit to add the two catch diodes and two snubber resistors to the back of your uTracer main board.

They are important to control potentially large ringing negative flyback voltages during a short circuit event.

The diodes should be high speed types such as 1N4937 OR UF4007.

The resistors can be any 1/4 1K  part.

[Walter Trovò]

Thank you for the feedback, Bob.

And thank you for the head-up on the extra protections: I will add them asap!

Walter

On Mar 13, 2025, at 22:53, Never Mind <bobblue...@gmail.com> wrote:



To view this discussion visit https://groups.google.com/d/msgid/utracer/f7c2ab88-a1f1-400d-8027-0d4e8ea11f7cn%40googlegroups.com.

<uTracer3 added diodes and resistors.jpg>