Current sensing

[Thomas Mueller]

I have an issue (I believe) with my uTracer6 and I was hoping to get some advice from the forum.  When testing my unit after construction, the anode and screen current/voltage plots are different.  It appears that the anode HV switch starts with elevate current compared to the screen.  Any idea's on how to troubleshoot this issue.  Also, when the range is set to "automatic" the anode current is capped.  See photo's below.

Thanks,

[Never Mind]

Tomm

A few notes on the uTracer6.

1)

Ronald's GUI has a bug in it during Va=VS sweeps that results in the incorrect screen voltage being reported under conditions where the screen current is different from the anode current by a large amount.

This bug is most evident on a uTracer6 below 100V and at higher currents. This may not be your present issue but did cause me to nearly toss my uTracer6 under a bus before I figured it out why low voltage high current measurement were so inaccurate.

uTracerJS does not have this bug.

2) The very low current measuring accuracy of a utracer6 is about 3.8 times lower than that of the utracer3+. So very low current measurements are never as accurate as on a utracer3

3) The low voltage measuring accuracy on a utracer6 is about 2.35 times worse than a utracer3+. 

So low voltage measurements loose some accuracy on the uTracer6 as there is only about one bit per volt resolution in the utracer6 AD for Va and Vs.

The chart you red circled "PIC2" is at both a low current and low voltage where the uTracer6 is weakest in accuracy. Is this your issue, I am not sure.

Trying a sweep with a lower resistor value such as 1K to increase the current may provide a hint if this is a low current measuring error or just a Va,Vs measuring error.

If it turns out to be errors in measuring Va and Vs at very low voltages it may not be a practical problem when measuring almost all tubes as results below 20V is normally a less important area.

The chart you red circled is also a Va=Vs sweep with Ronald's GUI screen voltage reporting bug. Is this a factor in you case, I am not sure?

You can do separate Va and then VS sweeps with Ronald GUI to see if the Va=Vs bug is part of your issue.

Trying uTracerJS will also tell you as it does not suffer from the Va=Vs bug.

uTacerJS has other features that Ronald's GUI does not that are useful in having a utracer6 reach it's best accuracy.

Finally it looks like the calibration accuracy of your uTracer6 could possibly be improved with a few more tweaks.

I suggest this as the VA curve is steadily higher than the Vs curve. If they were on top of each other in all areas but where the red circle is then the departure of the two curves at the bottom left would look a lot smaller.

BYW what is pic1 trying to tell me?

Happy hunting

[Never Mind]

I just got the point of Pic1.

The current range setting function in the PGA113 for the VA (IC62) is possibly not working.

Carefully check that pins 27 and 28 on the PIC (IC1) are fully inserted into the socket as they control the PGA113 select pins.

Try setting each current range manually for VA and VS to note what ranges the match and what ranges differ.

That will tell you if it is only a few PGD113 range settings that are not working or if range setting is not working at all.

This may provide clues as to what pins may not be properly soldered or shorted on the IC62 PGD113.

The pins are small so a low cost digital microscope I find helpful inspecting pins this size pin.

[Thomas Mueller]

bob,

Thanks for the reply.  Yes, pic1 is the main issue I'm seeing.  The anode current doesn't autorange.  Pins 27 and 28 on the PIC look fine and are seated.  I used a jewelers scope to look closely at the pins on IC62 and they all look ok.  I can replace IC62 with a new one.  I'm comfortable soldering small SMD parts.  Is there anything else I can check before I do so?  Could it be a floating ground issue?  Thank you for your help.

tom

[Davo]

Do you have a thermal cam or digital thermometer? Or just touch it..

And check the PGA’s temperature, if it’s getting really hot it’s toasted.., 

Op 14 okt 2024 om 06:32 heeft Thomas Mueller <tomm...@gmail.com> het volgende geschreven:

bob,

--
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 on the web visit https://groups.google.com/d/msgid/utracer/d843a90e-74b5-4cb1-92fb-a3b5540560bbn%40googlegroups.com.

[Never Mind]

Tom

The data sheet for the PGA113 shows the following.

"POR (Power-on-Reset) value of internal Gain/Channel Select Register is all 0s; this value sets Gain = 1, and Channel = VCAL/CH0"

The current measuring voltage buffer (IC7) is connected to channel 1 of the PGA113.

To me this means that if the PGA113 was not being programmed at all current measurements on the anode channel would be close to a zero value or the value of default channel zero.

That is not your case as setting the GUI to manual current mode the reading was fine on the lowest current range setting.

Did you try manual setting of all other current ranges to see what ranges on the anode channel match the screen and what ranges do not compare to the screen channel?

This may provide some useful clues before you start replacing parts.

It could be a defective part...

In my experience in a newly built products defects are most often assembly in nature, soldering opens/shorts, PCB track defects, incorrect part stuffing and so on.

I have often replaced the "defective" part debugging new builds only to find it made no difference and later finding the assembly defect.

It is possible a error during assembly did damage the anode PGA113 if you "slipped" with a probe at any point testing the utracer!

Such is life.

Best of luck.

BTW

After some struggle my utracer6 did work correctly so you should be able to get there.

[Thomas Mueller]

Bob,

I was able to run the full range of manual current settings.  I have attached pictures.  I don't know what to make of them.  The anode current is really odd though each current range.  I can almost see a pattern as I cycle through the pictures.  Can you think of what would cause this?  Could this still be an issue with a potentially defective PGA113?  I appreciate the help.

Tom

[Never Mind]

Tom

Interesting results.

I take the following from your results.

1) 

All PGA gains can be correctly set by the PIC.

This suggests the connections for Ci Di and -CS between the PIC and the PGA are fine.

It also suggests to me the PGA may not be the defect.

2) 

The current readings between the anode and screen on all ranges manually set are similar but not exactly the same.

The differences are small enough to possibly be a combination of calibration errors and component tolerance.

Hard to be sure at this point but it does not seem to suggest at this point any gross failure in the IC62 the PGA113.

3) 

The anode curve result for the manual setting 0-4mA is exactly the same as the auto  anode curve result.

This suggests the PIC is ALWAYS SETTING the anode PGA IC62 to maximum gain when set to auto.

I do not have the source code for the PIC as it is black box closed source code :(  so I can only speculate on how the code works.

I hate debugging black box micro code hardware faults but what can you do?

However I am going to suggest the following.

The PIC first measures the voltage level on AD AN0 (pin2) the input to IC62 the anode PGA.

Second based on what voltage it sees at the input to the PGA it then selects the correct gain range for the PGA.

The PIC next measures the value at the output of the PGA with AD AN2 (pin4). This is the value it presents in the results data. 

The PIC does no bounds checking for the value it gets at AN2 so if the data is full scale  (PGA clipping) you simply get a full scale reading of 4.5mA.

This makes sense as a good code strategy as it would always allow finding the correct PGA gain setting in a maximum of two AD measurements of the test pulse and would be consistently fast.

The lack of any bounds checking is common as programmers almost always assume the hardware is good and rarely sanity check results from hardware measurements. 

After all it takes more code, more real time and more programming effort to check your results for errors. Human factors are what they are.

4)

So my money is on there is a connectivity failure between PIC pin AN0 (pin2) and the PGA for the anode, IC62 CH1 (pin2).

This could be a PIC not fully set in the socket (common) OR pin #2 of the PIC is bent OR PIC pin#2 of the socket is not soldered OR there is a copper track defect on the PCB between PIC pin #2 and IC62 pin #2.

I suggest buzzing the connectivity out between the actual pin #2 on the pic and at actual pin #2 on the PGA. 

Look out in testing the connectivity as the connection may be intermittent if the cause is a loose connection. 

May times I have encountered testing shows all connections are fine because pushing on the pin with the ohm meter probe corrects the loose connection.

That is my best shot on you issue.

Good luck and let me know how it goes.

Bob

[Ihor]

If the the blackbox construction is of interest to someone in the future as well, here is the description what PIC does:

https://www.dos4ever.com/uTracerlog/tubetester2.html#more17

search for "(Auto)-Ranging" there.  

Probably indeed not a good connection, or some bad "interconnection" with something else that weakly pulls up the ADC input, forcing the Autoaveraging to default to high current range.

[Never Mind]

Ihor

Re:"Figure 14.4 Basic automatic gain selection algorithm"

You may be correct however the evidence from TOM suggests an alternate explanation.

Yes I have read that section however the hardware does not behave like it would if the code worked that way.

So I currently think perhaps the description on Ronald's WEB site is the "old" way and a simpler faster way was used at a later date that follows the description I gave.

If the code worked as described in Ronald's site it would not always select the highest gain range resulting in the PGA and the AD converter clipping.

As Tom can manually set the PGA gain and all PGA gain ranges work correctly when manually set (or very closely to the screen port) it seems clear that the anode PGA input is connected, the PGA output is connected, all PGA grounds and power are connected and the SPI to the PGA is working fine. 

There is nothing left that is needed for the PIC to "select" and set the correct PGA gain range if the code works as Ronald describes on his WEB site and only looks at the PGA output pin level.

To behave in the way Tom describes in his uTracer6 the PIC code must be dependent on some other signal than only the PGA output voltage to select the correct PGA gain range. 

Tom's PIC always selects the highest gain range no matter what the current signal on the output of the PGA is and that would not happen if the PIC was measuring only the PGA output voltage to determine the correct gain range.

So I am currently in the camp of  "do not believe everything" on Ronald's WEB site is how final uTracer PIC code was implemented. 

It may well have changed from the original uTracer3 code implementation and not have been documented on the WEB site.

 I may well be wrong, time will tell. I considered lifting PIC pin #2 on my uTracer6 to see what would happen but as it is currently working nicely I did not wish to temp fate.

Bob

.

[Thomas Mueller]

Thanks for that link.  I will review the auto-ranging logic.

[Thomas Mueller]

Thanks Bob for this detailed analysis.  What you describe makes sense to me.  This will give me some leads to chase down.  

I will let you know what I find.

Thanks,

Tom

[Thomas Mueller]

I think I found something.  When I was checking the traces between PIC (pin 2) and IC62 (pin 2) I have good continuity.  The pins on the PIC are intact and fully seated.  This appears to be ok so far.  I then measured voltage on IC62 (pin 5) which is the PGA output.  With the unit powered on, but not sweeping, I get ~250mv on pin 5.  When I measured pin 5 on IC82 I get 0v (~3ma) which I believe is correct.  

I think the 250mv on IC62 output is the error causing the current ranging problem.  When I sweep, IC62 pin 5 appears to operate normally as compared to IC82 pin 5.  It just never zero's out.  It stays at 250mv.  I figure it is a bad IC62 or a bad PIC.  Does this seem right or what else could cause the 250mv on pin 5 of IC62?

Thanks again.

Tom

[Never Mind]

Tom

Interesting.

Is seems my theory about the PIC auto-range code changing from how first described in Ronald's Web site is wrong.

Oh well theories are made to be tested and if required discarded. Next.

I assume you also checked the output voltage at rest of IC61 a OPA227 (PGA input on pin#2) and found it to be very close to zero volts.

Seems reasonable that if the PGA input is very close to zero and the PGA output is 250mV there is a problem in the PGA circuit.

As we do not have the PIC source code we do not know what the PIC sets the PGA gain to when the uTracer is at rest.

Keep in mind that the PGA gain can be set as high as 200, so a 250mV output from the PGA would require only 1.25mV output offset voltage from IC61.

IC61 is a OPA227 and the data sheet specifies the offset voltage as 75 μV maximum. In a differential circuit the output offset of IC61 the OPA227 should be less than about 150uV.

Best to check the output voltage levels carefully between the output of IC61 and IC81 if you have not already done so to make sure that they are both very close to the same and zero volts.

Ihor suggested some sort of crossed net connection to the PGA output and a nearby track may be dragging the PGA output off the correct voltage.

This may well be the issue.

There is also still the possible failure of the PGA113. Only way to be sure is changing it.

What I can not understand is why a 250mV DC offset on the PGA output pin at rest would cause the PIC code to stop auto-ranging during a test pulse. 

Wish we had the source code so we could better understand this failure mode.

Take care

Bob

[Thomas Mueller]

I confirmed the outputs of IC61 and IC81 are both 0v.  It must be the PGA113.  I ordered a new part a couple days ago to get it on the way in case I was to replace it.  I think it arrives today.  I will replace this weekend and respond back to this thread.  Thanks for the help on this.

Tom

[Thomas Mueller]

Swapping out IC62 (PGA113) worked!  It was bad.  See attached plots.  Thanks for your help.  I don't think I could have troubleshoot this by myself.

Tom

[Never Mind]

Tom

Great news.

As a bonus we both now know more about how a uTracer works.

It is a interesting device.

Bob

[Davo]

Just being curious, but did you checked if the PGA got hot?

I use a thermal cam quite a-lot when a device is broken to detect shorts ect.

A while ago I burned my uTracer3 when measuring the grid voltage (40v) on R4, and I accidentally shot out at R3, creating a short between R3&4 

Which shorted out the whole 5v line.

Bye bye PIC, 2 PGA’s, 2 BC’s LM & T3

With the thermal cam it was quite an easy fix

It was quite a stupid fault, but an expensive one to repair.

Op 20 okt 2024 om 18:19 heeft Never Mind <bobblue...@gmail.com> het volgende geschreven:



To view this discussion on the web visit https://groups.google.com/d/msgid/utracer/4f2e390c-91ad-4a69-b407-348bd279c94fn%40googlegroups.com.

[Thomas Mueller]

It didn't get that hot to the touch.  At one point, I was looking into thermal cameras, but I never purchased one.  I can see where they are useful when troubleshooting issue.

[D.A.R Achterberg]

If you can you can touch it it’s ok!

If shorted they can get 90/100 degrees Celsius in 10 seconds or so

  

Op 26 okt 2024 om 03:03 heeft Thomas Mueller <tomm...@gmail.com> het volgende geschreven:

It didn't get that hot to the touch.  At one point, I was looking into thermal cameras, but I never purchased one.  I can see where they are useful when troubleshooting issue.

To view this discussion visit https://groups.google.com/d/msgid/utracer/52f924c3-a738-48d0-9653-9982ad2300c3n%40googlegroups.com.