Super easy wall adapter mod: 9V plus 170V at 80mA

[taylorjpt]

Just for fun I modified a Cincon TR1509 wall adapter (9V at 1.4A rated) so that now it puts out the original 9V plus a regulated 170V at 75mA (13W total for both outputs).  This requires re-winding the transformer and then adding 2 diodes and a capacitor.  The original secondary is 2 x 15T of  #24 in parallel and is right on the outside of the bobbin so it is easy to get to.  The transformer core has a center pole gap so it is easy to take apart because only the outer legs are glued; I took this one apart with a BIC lighter.

The new winding is 15T #32 to replace the original winding followed by 268T #32 in series with the first winding.  The 15T winding goes back into the holes for the old winding and the new fly lead connects to a couple of high speed diodes that feed a high voltage electrolytic cap.

The main output is regulated to 9V with the original feedback circuits and the new winding puts out a voltage that is turns ratio related to the first.

One big advantage is that you don't get the double efficiency hit of an AC-DC feeding a DC-DC boost since all of the power is converted in a single step off the mains.  Also, you get the same mains isolation as the original design because you never have to touch the primary... most flyback converters sandwich the secondary between two halves of the primary for better coupling.

jt

[gregebert]

Clever!  I admire your patience winding 32 gauge wire; it's pretty fragile.

[taylorjpt]

My "Thin" wire is #43 or #50. This stuff feels like wire brush bristles when I use it.

Running a long time test: 185V at 75mA plus 9V at 100mA. Diode temperature rise is 8C and ripple on HV is 78mV RMS. Going to let it run til Saturday morning (When I need my bench back).

[taylorjpt]

Just modified a second unit... so easy.   By adjusting the output winding turns I should be able to get just about any voltage I want, i.e. 300V for a dekatron project etc. I replaced the 2 conductor cable with a 3 conductor cable that can handle the high voltage.

Anybody want this unit?  On to the next project...

jt

[David Forbes]

John,

That's a clever hack!

Once thing... you had to saw apart the ultrasonic weld to get it apart,
right? Any good ideas on gluing it back together?

Also, this is not strictly still UL recognized since you modified it,
but it's close enough. Do you consider it to be as safe as the original
unmodified supply?

On 10/13/15 5:35 PM, taylorjpt wrote:
> Just for fun I modified a Cincon TR1509 wall adapter (9V at 1.4A rated) so
> that now it puts out the original 9V plus a regulated 170V at 75mA (13W
> total for both outputs). This requires re-winding the transformer and then
> adding 2 diodes and a capacitor.
>

> jt
>


--
David Forbes, Tucson AZ

[gregebert]

I'm curious about adding windings to the transformer. Is this particular transformer constructed so you can pop-out the bobbin, or does it have those obnoxious interleaved E-cores that require you to split them all apart, recoat, etc ? Or worse yet, thread each turn thru the core ?

[taylorjpt]

No sawing.  You shove a flat blade screwdriver between the output cable strain relief and the cover and pry it up.  This breaks the weld which you then work around until the lid pops off.  Very little plastic deformation which then glues back together in a snap after the mod.

Once thing... you had to saw apart the ultrasonic weld to get it apart,
right? Any good ideas on gluing it back together? 

Never contended that this was still UL recognized after the fact... My statement was simply that the mains isolation was maintained because of the construction of the transformer.  For transformers that have split primaries above and below the secondary, I would be reluctant to use them after modification in any case.  Where the split primary linear power thread uses transformers outside of their intended application, this modification mantains all of the original creapage distances.
 

[taylorjpt]

High frequency switchers use ferrite EE cores.  You cant really see it from the picture but the center posts are slightly shorter than the legs (About 0.015" total) and they are glued together with a drop of epoxy on the legs to hold them together.  The reason that they gap core is to lower the inductance (Permeability of the magnetic path) which increases the saturation current and thus the energy storage capability.  The reason they gap the center post is so that any leakage field is contained within the windings and faraday shielding to prevent radiated interference.  Another benefit of the air gap is that slight errors in the mating surfaces have little effect on the inductance so long as those errors are a small percentage of the gap distance.

1.  Cut the epoxy drops off the outer legs.
2.  Heat the cores with a heat gut until the varnish is softened (an exacto knife pokes it easily).
3.  Carefully pry the core on the top away from the bobbin.
4.  Push out the bottom core with the center post
5.  Clean up the excess varnish and polish the legs in preparation for re-assembly.
6.  After re-winding the transformer, use a couple of drops of super glue on the outer legs to glue them back together.

[David Forbes]

All high frequency switching transformer cores are cast ferrite. No
laminations to deal with. They are called E cores because they look like
the letter E.

John said that after he removed the transformer from the board, he
separated the ferrite core halves by heating the glued joints with a Bic
lighter, then pulling the cores out. Then you unwind the tape form the
bobbin and remove the outer winding.

[taylorjpt]

A more elegant method for adding the 3 conductor cable (Than that showed in my photo) is to take the original strain relief and pull out the wire and then drill it out.  The new wire is then slipped into the strain relief.

[taylorjpt]

Here's a closeup of the re-sealed housing with the re-fitted strain relief.

[taylorjpt]

I did the same process on an HDP24-1AWA01FZ (12V @ 2A) of which I have a couple hundred scrap units and am getting a clean 170V at 140mA.  These units are welded better so I should find an easier way to open them.

This is so much easier than the way I tried to do it years ago by trying to redesign the feedback circuits to make the switcher regulate the high voltage directly:  I still needed a low voltage rail to generate the logic bias for the opto-coupled feedback circuits which meant adding a linear post regulator or a separate winding and output rectifier/filter.  By allowing the switcher to run at its designed output voltage and putting up with sloppier load regulation I am getting the job done with no board modifications.

[John Rehwinkel]

> This is so much easier than the way I tried to do it years ago by trying to redesign the feedback circuits to make the switcher regulate the high voltage directly: I still needed a low voltage rail to generate the logic bias for the opto-coupled feedback circuits which meant adding a linear post regulator or a separate winding and output rectifier/filter. By allowing the switcher to run at its designed output voltage and putting up with sloppier load regulation I am getting the job done with no board modifications.

This is really clever. I wish I had a transformer winder!

- John

[Dekatron42]

Have you tried the same process of rewinding the transformer on any of the small switching power supplies meant for driving LEDs, like the one shown in the picture below? The market is currently swamped with these power supplies and they are very cheap and modifying one or two in addition to using them as they are seems very attractive to me at least.

/Martin

[gregebert]

I'd say it's dependent on :

1. How much core-area is available for the additional winding

2. How easy it is to remove & reinstall the bobbin

3. The total power of the supply

If you're going to drive six IN-18's (the largest commonly-available nixie) at 150VDC, that's about 5 watts. Any power used for the new winding would need to be offset by using less power from the original output(s). Otherwise you will be overloading the supply.

[taylorjpt]

You are better off using the highest main output voltage you can tolerate because of the relatively poor coupling of the low and high voltage windings that results in an error in the un-regulated output (In this case the 170V for instance):  This error is manifested as a change in the high voltage output with a change in load on the low voltage output as well as the load regulation of the high voltage output itself.  The lower the difference between the regulated and unregulated voltages, the better the winding coupling and therefor the better the cross regulation.  The high voltage output operates like a peak detector when un-loaded so even a little bit of pre-load (Say 1%) goes a long way to improving the load regulation of the high voltage output.

1.  The winding area (Not core area) for the secondary of these transformers is designed to support the output POWER for which the power supplies are rated so there is ALWAYS enough winding area, regardless of the voltage that is output.  Few thick wire turns for high current == Many thin wire turns for low current.

2.  The Cincon TR1509 is a perfect platform for this modification as the secondary is the outermost winding and the cores are glued in a very easy way with regard to separation.  The varnish used in the industry has a relatively low transition temperature:  I am using a Saike hot air reflow station set to 425F and I heat the core uniformly until the varnish between the core and bobbin yields with gentle pressure.  Avoid applying heat to the entire transformer vs just the cores.

3.  What ever the power supply is rated for in watts, that is the power that can be delivered before and after the modification.  Note that most power supplies have two current limit mechanisms;  Primary side and secondary side.  Adding a second winding will still use the primary side limits as the current to deliver a particular power output remains constant but since the currents are much lower at higher output voltages for a given power level, any secondary current sensing will never come into play.  However, any secondary over voltage protection will still be active since the regulated output voltage has not changed.  The 9V output of the Cincon TR1509 is just right as it gives plenty of headroom for a 5V regulator while maintaining a reasonable primary to secondary turns ratio.

It would be possible to feed the high voltage into the feedback network while using the low voltage rail to bias the opto-coupler circuits.  This would result in a tightly regulated high voltage output which sloppy low voltage regulation but with sufficient overhead for a 5V regulator the application circuit would not see the variations.  Although this goes a bit further than most people are comfortable with, this would have the added benefit of preventing load variation or multiplex scan induced brightness variation in the circuit.  I am getting a PWB made that mounts on the HV capacitor to hold the HV diodes and a voltage divider just for this improved scenario.

[gregebert]

If you drive your tubes with a current-regulator, there's really no need to regulate the HV supply unless you are trying to minimize the power. All you need is to ensure the voltage is high enough for ionization.

I've used current-regulated cathode drivers on two of my clock designs. My present clock project uses a current-regulator on the anode side (each tube has it's own PMOS device). The drive current is adjustable with a single pot; I didn't see a need to tweak each tube's current independently, though that is possible by swapping a resistor in the current-regulator.

I've chosen current regulation instead of voltage regulation because the brightness and 'wearout' of the tube is a function of the current. Nixies have nonlinear I-V characteristics, and I've seen considerable variations between tubes. All of this matters when you use a constant voltage + series anode resistor to run the tube because you have to calculate the resistor value based on min & max supply voltage, min & max tube-voltage, etc. It's moot with a current-regulated driver.

[Steven Donaldson]

Are you willing to post a schematic of your current regulator driving multiple tubes?

Thanks,

Steve

--
You received this message because you are subscribed to the Google Groups "neonixie-l" group.
To unsubscribe from this group and stop receiving emails from it, send an email to neonixie-l+...@googlegroups.com.
To post to this group, send email to neoni...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/6d72ff6d-d737-4316-a26f-f7012f999f70%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[taylorjpt]

The Cincon TR1509 is actually quite amenable to directly regulating the high voltage output directly by removing R22/C4 and then feeding the top of R21 with the output of a 511K/4.7nF tied to the high voltage rail for a regulated 173V.  This has the advantage of not requiring any preload so long as the lower voltage rail is high enough to run the application circuits.

Current sinks are the easiest thing to build so long as you have some reference voltage to start with.  The PDF shows a simple saturated cathode drive for well regulated HV and two variations of current sinks (Actually a current sink and current source!) where the first needs to be attached to every cathode and the second can be shared between cathodes in a multi cathode display such as a 1-of-10 nixie.  The current sinks have the advantage that they are relatively immune to variations in the high voltage rail as well as variations in cathode conduction voltage.

[taylorjpt]

Would anyone like one of these modified units to play with? I am getting 300 meters of good 3 conductor high voltage cable and plan to terminate it with a 4 pin 1.25mm JST connector. These will be the high voltage regulated versions.

[taylorjpt]

Using the high voltage output as the regulated rail really works well!  The low voltage output only varies by about 15% for a high voltage load from 0 to 75mA (The 170V was rock solid).  A few more benefits from this implementation, and this mod in general:

1.  No boost converter heating in the target circuit as all the conversion is done in the wall adapter.
2.  Super quiet operation, audio and electrical since the switcher operates exactly the same whether it outputs 9V or 170V.
3.  Putting the feedback on the high voltage serves as a bleeder current to discharge the output capacitor.
4.  Scalable to other voltages:  I did a test with another turns ratio and am getting a clean 14W at 300V.
5.  Scalable to higher power levels:  I did the same mod on a 24W unit and got 140mA at 170V.
6.  Failure of the high voltage supply can be fixed by swapping the adapter instead of opening the target circuit.

[A. Nonamus]

I think there's a small typo in Current Sink.pdf. At the end of the description for the circuit on the left, "(Ir1 / 100) to satisfy Q1 hFE.", shouldn't that be Ir2?

[taylorjpt]

Thanks, I should have spent more than 15 minutes on it!  Here is an update, hopefully I didn't add more errors than I fixed.

[taylorjpt]

Most nixie controller circuits don't use that much power, most of the power is used from the high voltage to light the displays. This modified wall adapter technique actually delivers more usable power for a given adapter power because the adapter efficieny is about 80% but if you take the low voltage and boost it up to the high voltage, that converter is also about 80% efficient. This means that the nixie power is delivered at 80% x 80% or 64% efficiency.

[taylorjpt]

I was talking to Cincon and they might actually be able to either sell me unwelded units for modification or even make a custom unit. 

How many people would be interested in an AC adapter that puts out 9V and 170V at 15W total?  These could also be done in a 9V/300V output version.

jt

[Dylan Distasio]

I would probably be interested in both, thanks.

--

You received this message because you are subscribed to the Google Groups "neonixie-l" group.
To unsubscribe from this group and stop receiving emails from it, send an email to neonixie-l+...@googlegroups.com.
To post to this group, send email to neoni...@googlegroups.com.

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/6aa437e0-28a9-472b-8cb4-3127f4d80d27%40googlegroups.com.

[Jon D.]

I'd be interested in a couple of each.

Jon

[John Rehwinkel]

> How many people would be interested in an AC adapter that puts out 9V and 170V at 15W total? These could also be done in a 9V/300V output version.

While your existing supplies are very handy, an all-in one would be a nice solution too. How would it be terminated? Bare wires? TRS connector?
I'd be interested in buying a few. You could maybe even offer a variant with 6.3V for tube heaters and B+ as well. Granted, the tube heater would
probably eat most of your wattage, but for a headphone amp or phono preamp, it would be a nice fit.

- John

[taylorjpt]

Currently I'm planning to terminate the 3 conductor cable with a 4 pin 1.25mm JST connector: 9V/Gnd/NC/HV.

Jt

[Dekatron42]

Can you supply them with the Swedish (European 2 pin version) plug too?

I've tried to buy these since you started this thread but every reseller has told me these are obsolete, like it says in the link from Mouser here: http://www.mouser.se/ProductDetail/Cincon/TR1509-A-12A03/?qs=Pauzw2weiyNd0vxdPKjLgA%3d%3d

I'd be interested in two each of the 170V version as well as the 300V version.

/Martin

[Donald Stramock]

I would be interested in the 9V plus 179V at 80 ma

/DonS

Sent from my iPad

--

You received this message because you are subscribed to the Google Groups "neonixie-l" group.
To unsubscribe from this group and stop receiving emails from it, send an email to neonixie-l+...@googlegroups.com.
To post to this group, send email to neoni...@googlegroups.com.

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/3aa8f2bd-e477-4737-9382-dc39f4ececf7%40googlegroups.com.

[GeckospotNixie]

I would like 6 Please

[taylorjpt]

They have been replaced by the newer tech TRG1509. The transformer is smaller as it runs at a higher frequency but the new windings still fit due to the adjusted turns ratio:

Model 9V 170V
TR1509 15T 250T
TRG1509 11T 184T

The TRG1509 adds a faraday shield to the outside of the transformer but it does not impact the rewinding of the secondary which is the outermost winding in both models.

[Dekatron42]

Thanks for pointing out the new model, unfortunately I can only find that one with Mouser but Mouser says: "Due to government regulations, Mouser is unable to sell this product in your country." , so I would still need your help in getting hold of the type with European (Swedish) prongs.

/Martin

[Dekatron42]

I decided to buy a few Mean Well switching power supplies the other day as I found them on sale for SEK 20kr each (some USD $2) to see if I could modify them - I decided to buy them because they were cheap and I could use them for other stuff if I couldn't rewind them.

I have successfully modified a RID-50B (5V@4A and 2...@1.4A) (http://www.meanwell.com/mw_search/RID-50/RID-50-spec.pdf) to get 5...@1.5A and 198V@60mA, but it was more complicated as the secondaries were put in between the primaries, but it was quite easy to modify anyway. Now I need to go and buy some more transformer tape as I ran out of mine. It was easy to remove the outer primary winding and I could reuse most of the tape that I found inside the transformer, even the outer tape was reusable. I did remove part of the 5V winding and didn't put it back so I only got part of the original current, it was wound both underneath and on top of the 24V winding that I replaced. Without load I get some 210V and with 3.3 KOhm load I get 198V.

It was easy to modify the core as it was not glued together at all, just kept in place by some turns of transformer tape on the outside and then some waxy substance in the middle of the core, the sides were not glued at all! Adding 140 turns of 0.25mm enamelled copper wire was quite easy and it fit fine on the core with enough layers of transformer tape to keep the isolation intact (I put on two more layers to be sure to not get any isolation problems but I am not sure if it affects the transformer negatively in any other way. I only had UF4007s (probably not the best ones but they work) to use as rectifiers and those together with a 22uF/400V capacitor was enough to keep the hum down below 150mV, I haven't added any other components to the output for now but it looks promising.

Thank you John for publishing your modification, it made me brave enough to take the step to try it myself! Now I just wish I could find those wall warts or anything similar here in Sweden.

What rectifiers did you use?

/Martin

[taylorjpt]

I'm using two STTH1R06A in series. The switchers usually operate between 50 an 100kHz so you want to keep the reverse recovery time below 50nS. 25nS or better at the low currents used is pretty easy to find. Just as important is the reverse voltage rating as these need to support the difference betwee the output voltage and the negative voltage during the primary on time which is worst at maximum input voltage.

[taylorjpt]

My question was primarilly to see if there was any interest in a comercial version of this... however I will be making up quite a few as proof of concept that will be given away after testing is complete.

[croftj]

You may want to make it configurable or be willing to sell them so that we can make them what we want. I personally think a 90 - 100v version would be sweet!

-joe

-------- Original message --------
From: taylorjpt <j...@tayloredge.com>
Date: 23/10/2015 6:42 PM (GMT-06:00)
To: neonixie-l <neoni...@googlegroups.com>
Subject: Re: [neonixie-l] Super easy wall adapter mod: 9V plus 170V at 80mA

I was talking to Cincon and they might actually be able to either sell me unwelded units for modification or even make a custom unit. 

How many people would be interested in an AC adapter that puts out 9V and 170V at 15W total?  These could also be done in a 9V/300V output version.

jt

--
You received this message because you are subscribed to the Google Groups "neonixie-l" group.
To unsubscribe from this group and stop receiving emails from it, send an email to neonixie-l+...@googlegroups.com.
To post to this group, send email to neoni...@googlegroups.com.

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/6aa437e0-28a9-472b-8cb4-3127f4d80d27%40googlegroups.com.

[Charles MacDonald]

On 15-10-24 08:33 PM, croftj wrote:
> You may want to make it configurable or be willing to sell them so that
> we can make them what we want. I personally think a 90 - 100v version
> would be sweet!

this sort of thing would fit a few niches. I am not sure if a little
switcher would handle the inrush current of a Vacuum tube filament but
something that gave enough power to light up a couple of 12AX7 tubes or
something similar with several ma at 170 volts would be quite
interesting for both the audio and Guitar camps.

(12AX7 Draws 12.6 volts at 150ma or 6.3 volts at 300ma.) and likes
anywhere from 100 to 250 volts on it's plate.


--
Charles MacDonald Stittsville Ontario
cm...@zeusprune.ca Just Beyond the Fringe
No Microsoft Products were used in sending this e-mail.

[taylorjpt]

These switchers can deliver the full output power from any combination of output loads so long as the feedback circuits are correct. The primary side controller would not know the difference since it stores the same energy in the transformer during the on time regardless of how it is output during the off time. For vacuum tube filaments a simple push/pull switcher would do a nice job and could deal with the lower cold filament resistance.

Since the output voltage is turns ratio dependent, it would be easier to wind them as required for specific output vonfigurations. For adjustability of the HV output, the low voltage output would be variable within a useful range. For instance, for a 100V to 200V output range, the the HV winding is calculated for the highest voltage and then the low voltage winding would be calculated to give the minimum usable voltage for the minimum HV output.

[taylorjpt]

That's what I like about the Cincon TR(G)1509: The chose to not split the primary. The currently avaliable G version still did not split the primary and only added a Faraday sheild winding outside the magnetic circuit (Around the outside of the core) which is easy to deal with.

Every other switcher I've pulled apart splits the primary. The main reason to do this is to limit the leakage inductance: Magnetic field of the primary curcuit that is not part of the secondary curcuit. Leakage inductance shows up as a voltage spike on top of the primary winding added to the turns ratio voltage when the primary transistor turns off, kind of like an ignition coil. You either have to way over rate the transistor or add an energy wasting snubber to keep the transistor from dying.

[Dekatron42]

Thanks for the explanation and for what rectifiers to use, could only find them in spools of 5000 pieces in Sweden so I'll have to go shopping at Mouser or Digikey.

In my first attempts I found out that just one UF4007 was not enough to get more than some 20mA before it broke down so I looked at your first photo and saw two rectifiers in series there (at least I thought it was two in series) and tried it and it worked a lot better! Some Googling also showed this use of two in series. It says 50 kHz switching frequency in the datasheet for the RID-50B, so I guess that I'll get by with the UF4007s until I get hold of the STTH1R06A.

I'll make one change when I have some more time and that is to move the second primary winding down so that the high voltage winding will be on the outside.

Mean Well uses snubbers to get rid of the spike you mention.

/Martin

[Bill van Dijk]

John,

Although tempting, you may want to rethink that. Especially in a pre-amp the switching noise may be a problem.

Bill

-----Original Message-----
From: neoni...@googlegroups.com [mailto:neoni...@googlegroups.com] On Behalf Of John Rehwinkel
Sent: Friday, October 23, 2015 9:59 PM
To: neoni...@googlegroups.com
Subject: Re: [neonixie-l] Super easy wall adapter mod: 9V plus 170V at 80mA

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

To post to this group, send an email to neoni...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/B57A82A7-FB1C-4CDB-9663-98D972C8886D%40mac.com.

For more options, visit https://groups.google.com/d/optout.

---
This email has been checked for viruses by Avast antivirus software.
https://www.avast.com/antivirus

[taylorjpt]

Yes, I will keep rethinking it until I make it work! That's why I have 18 patents.

[John Rehwinkel]

> Although tempting, you may want to rethink that. Especially in a pre-amp the switching noise may be a problem.

Possibly not. Switching is at a higher frequency than a mains transformer, which makes it easier to filter out. The usual issue I run into is radiated, inductively coupled, and capacitively coupled switcher noise getting into sensitive circuits. That's why this solution is appealing: all the switching circuitry is far away in a wall wart and only DC(ish) power is brought away from it. I'm guessing that with a little care on the power input, I could banish the switching noise quite well.

Since JT is offering to let people try units out, I may be able to test this directly.

- John

[taylorjpt]

The HV active preload boards are on their way!  This circuit keeps the high voltage (HV) output loaded so that the low voltage (LV) output does not collapse when the HV load goes to zero quickly, but also does not subtract from the deliverable output power of the supply.

To maintain tight regulation of the HV output I am dividing down the HV to 2.5V and injecting this into the feedback network, where the LV output is used to power the feedback comparator, opto-coupler and any user circuits.  The main advantage of this arrangement is that changes in the HV load do not show up as display tube brightness fluctuations (Or punch through in the case of audio amps).  An issue that arises however is that by using the high voltage as the feedback source with no preload, when the HV load abruptly goes to zero the switcher will skip cycles until the HV output capacitor is discharged to the regulation threshold resulting in the LV output momentarily collapsing.  Once the HV output recovers from the load step, the LV output will stabilize.  Although this is not a problem for the switcher itself, any circuit that uses the LV output may brown out during the recovery time.

The test condition for the AC adapter with this circuit is that the LV output maintains sufficient head room for the users LV regulator for an instantaneous step change in HV current from maximum (In this case 170V at 75mA) to zero.

[Dekatron42]

What happens if you short the HV output to the +9V output or to ground? Will both outputs survive and how will it regulate during the short and will it recover when the short is removed?

/Martin

[taylorjpt]

Shorting the HV output to the 9V output will destroy the 9V output circuits: Rectifier, reference, opto coupler etc as well as probably the user circuit (Unless it is protected). This can't be avoided because of the energy stored in the HV capacitor. This of course would be the same situation if the output of the HV converter in a clock was internally shorted to the low voltage rail in a clock, etc.

Shorting the HV to ground however has the same effect of shorting the LV output to ground: The switcher primary current sense circuit will turn the unit off. These power supplies are designed to handle this type of failure indefinitely.

My choice of a 5 pin keyed connector is with the ground at the center left was made to prevent the HV from accidentally being connected to the LV output even if forced in backwards and give extra isolation from HV to GND.

1 NC
2 LV
3 GND
4 NC
5 HV

[Dekatron42]

Thanks for the answer!

Nice with the keyed contact.

/Martin

[Dekatron42]

Would a TranZorb or something similar on the LV output be able to cope with a short from the HV output?

/Martin

On Sunday, 1 November 2015 12:09:13 UTC+1, taylorjpt wrote:

[Bill van Dijk]

I often use (flat) non-keyed connectors, in which case rearranging the connections slightly will make the connector non-polarized:

1 HV
2 LV
3 GND
4 LV
5 HV

Bill

-----Original Message-----
From: neoni...@googlegroups.com [mailto:neoni...@googlegroups.com] On Behalf Of taylorjpt
Sent: Sunday, November 01, 2015 6:09 AM
To: neonixie-l
Subject: Re: [neonixie-l] Super easy wall adapter mod: 9V plus 170V at 80mA

--
You received this message because you are subscribed to the Google Groups "neonixie-l" group.
To unsubscribe from this group and stop receiving emails from it, send an email to neonixie-l+...@googlegroups.com.
To post to this group, send an email to neoni...@googlegroups.com.

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/0315f3b5-7e03-477f-8941-2dfabe85082a%40googlegroups.com.

[taylorjpt]

Preload boards are in and final testing has begun.  This first unit is 9V at 0-100mA and 170V at 0 to 82mA.

[Dekatron42]

Looks nice!

I'm still waiting for some spools of transformer tape, 3M 44 margin tape and also 3M 56 polyester tape, I received a bunch of STTH1R06 diodes (the through hole variant) so I can continue experimenting soon.

/Martin

[taylorjpt]

Load testing is done and the transient performance is right on target using a 1.3W active preload. 

Plot on first page of 0214 PDF is voltage at the end of the new 2 meter cable 3 conductor cable with a 100mA 9V load and 0-82mA 170V load step.

Plot on the second page is the input to the HV rectifier showing the reverse bias stress:  As is with the 2 x 600V diodes in series there is 300V of margin at 240VAC input so for 240VAC in, the 800V or 1KV diodes of the same family would result in increased reliability and transient margin. The switching speed of the converter is only 31kHz (32uS period) so that the increase in Trr of the higher voltage devices will not be a problem.

The next step is to back the preload current off to its minimum while maintaining the 9V output cross regulation performance.

[Dekatron42]

Nice!

How would three or four STTH1R06 HV rectifiers work instead of two of the 800V or 1KV versions, just as well or would other problems arise?

/Martin

[taylorjpt]

3 would work just fine at 600x3=1800V vs 800x2=1600. I will switch to the 1KV parts since I already have PWBs.

The forward bias efficiency impact will be minimal for two vs three diodes as the forard voltage at these low currents is so much smaller than the output voltage.

It may even be a wash as the 600V parts are faster and energy is dissipated as the devices switch from on to off during those Trr nano seconds. A faster part is less energy. In either case, it's not worth the energy to calculate it!

[Dekatron42]

Thank you for the answer!

I realise I need better probes for my scope to make the measurement you did in the second scope picture, I only have standard probes that aren't rated for those pulse voltages. Where did you hook up the probe to show that trace?

/Martin

[taylorjpt]

The probe was connected right at the new HV winding where it enters the rectifier pair.

[taylorjpt]

 If you look at the second plot, the reverse recovery is not even a factor since the forward current goes to zero 10's of microseconds before the next primary on-cycle as evidenced by the secondary ringing signal.  The ringing also indicates something about the diode performance:  The cycle to cycle decay is almost zero indicating there are almost no losses in the resonant tank formed by the primary and secondary circuits parasitics.

[taylorjpt]

I sent the first fully integrated unit out a few weeks ago (can't remember to whom) so when I get some feedback I've got 10 more that I can finish up and send out to the first 10 emails, starting NOW; 1 per person, 170V and 9V output at 15W total.

Jt

[Jon D.]

I'm interested in one. Thanks in advance.

Jon

[Greg Jeziorski]

I am interested in one.

Thank You!

Greg

[Jon]

If that's still open, I'd be interested!

Thanks,

Jon.

[Jonathan Peakall]

I'm interested!

Jonathan

[Billy Watson]

I would like to test one and give a review.  Thanks

--
You received this message because you are subscribed to the Google Groups "neonixie-l" group.
To unsubscribe from this group and stop receiving emails from it, send an email to neonixie-l+...@googlegroups.com.
To post to this group, send an email to neoni...@googlegroups.com.

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/5664B362.1060003%40madlabs.info.

[Marcin Adamski]

I am interested too!
Cheers, Marcin

[taylorjpt]

Ok, these will go out in order of the requests here startin tuesday 12/15 (in maui till then).

Jt

[GastonP]

I'm interested, too!

[Keith Moore]

I assume I am too late for one?