SMPS for tube pedals

[antonis]

I cut the fins off the heatsink to save some room.

https://www.heatsinkcalculator.com/heat-sink-size-calculator.html
https://celsiainc.com/resources/calculators/heat-sink-size-calculator/
https://www.allaboutcircuits.com/tools/heat-sink-calculator/
]https://myheatsinks.com/calculate/plate-fin-heat-sink-calculator/[url]
https://coolingsourcethermal.com/heatink-size-calculator/[/url]

[Rob Strand]

LOL, I hope it didn't need those fins to dissipate heat well enough 

Possibly a sign the converter is a much higher power unit.
Higher power could mean higher current pulses on the input supply => more noise!

See below.

Ok here it goes - this has been tested and runs 1 .. 2 preamp tubes nicely.

Thanks for posting.

Interesting the timing cap value is 390pF.   The trace I had for the through-hole version quoted 100pF to 270pF
which seemed a bit low.  The 390pF (say 330pF to 470pF) seems more likely.

The SMD board has an SMD inductor.   From what I can work out it's 12mm x 12mm x 8mm high.   Wurth sometimes have different inductor heights for the same footprint and the inductors have different specs.
From what I can work out it's a,

Wurth (WE), PD SMT Shielded Power Inductor, 744770222
Isat   1.49A
RDC   247m ohm typ.,  390m ohm max

Here's the through-hole trace from the web with a few fixes,



There's no heatsink on the MOSFET.

Specs for OP's posted pic, aliexpress/ebay unit with heatsink.

DC-DC Boost Converter 8~32V to 45~390V High Voltage ZVS Step up Booster Module

Description:
High Voltage Boost Module
Module Properties: Non-isolated step-up module
Input Voltage: 8~32V input(the default is 10~32V input.)
Input Current: 5A (Max)
Quiescent current: 15mA (12V liter 50V, the output voltage, the higher the current will increase too quiet)
Output Voltage: +45~390V continuously adjustable (default output ±50V)
Output Current: 0.2A Max(with input, output pressure related,the higher the output voltage, output current is smaller)
Output Power: 40W (Peak 70W)
Working Temperature: -40 ~ + 85 degrees (ambient temperature is too high, please enhance heat dissipation)
Operating frequency: 75 KHz
Conversion efficiency: up to 88% (efficiency and input and output voltage, current, pressure-related)
Short circuit protection: Yes.
Over current protection: Yes. (Input current exceeds 4.5A, reducing the output voltage)
Over voltage protection: Yes. (Output voltage exceeds 410V, lowering the output voltage)
Input reverse polarity protection: Yes (non-self-healing, reverse burning fuse, try not reversed.)
Installation: Four 3 mm screws
Wiring: free welding output terminals
Size(L*W*H):60 x 50 x 20 mm

Comments:
- 40W much higher than load from tube circuits.   Will the module work with the light load of the tube ckt?
- "ZVS Step up Booster Module" - perhaps this will keep noise down.  Better than common MC34063/NE555

[snailspacejase]

Good stuff guys.
So if I were to do a vero layout of that Kingsley through hole circuit, is the priority to keep the traces around Q1, L1 and D2 as short as possible?
I'll draw it up and get you guys to critique it if you would be so kind?

[shawnee]

Gosh, this is all thoroughly wonderful!
I've been away from here (DIY stomp) for at least 15 years, did other stuff (art and then a few years of meditation!), but it's come back to haunt me, still don't know bugger all but I LOVE constructing!
But this thread is totally up my alley as tubes have always enthralled me and recently got a whole load of Chinese 6j1's, small signal pentodes, 7 pins as well as a good load of the Russian 6n2p eb (military 12ax7 pretty much).
Big part of that was coming across a guy in Poland, alex.hotkits@gmail.com, through facebook pedal groups, whom has pedal pcb's aplenty and, especially, gets paid through paypal then the boards (at least five) come direct from JLCPCB... I absolutely love that post modern car and share ideal!
I've a image share somewhere so I'll try and get that up and goin' as, with what's here as the boost converters theres one particular Seymour Duncan thing I've got to get goin'. SD sfx-04 twin tube mayhem (just had a quick look but couldn't find the schematic to link to but I do have a copy) and it's neato as it has both the HV requirements, for the pentodes as well us HV transistors (MPSA42 and MPSA92 set up at the input alike a fuzz face) and then a whole swathe of gyrated opamps on 15+/15- like the HM2 stuff. The sfx has the NOS tiny wee 6205 pentodes whereas I wanna use the 6J1's and the MPSA92 (PNP) cannot be found, cheaply at any rate, whereas the MPSA42 can, be found, and tipped upside down! 

[crane]

Thanks for posting.

Interesting the timing cap value is 390pF.   The trace I had for the through-hole version quoted 100pF to 270pF
which seemed a bit low.  The 390pF (say 330pF to 470pF) seems more likely.

The SMD board has an SMD inductor.   From what I can work out it's 12mm x 12mm x 8mm high.   Wurth sometimes have different inductor heights for the same footprint and the inductors have different specs.
From what I can work out it's a,

Wurth (WE), PD SMT Shielded Power Inductor, 744770222
Isat   1.49A
RDC   247m ohm typ.,  390m ohm max

Here's the through-hole trace from the web with a few fixes,

Yes - I used Wurth inductor for my build - but some other brand inductors where availible in my local "radioshack" store as well.
BTW - your schematic shows a wrong fuse value.

[Rob Strand]

BTW - your schematic shows a wrong fuse value

That's what was on the web schematic.  I suspected it was misread as 500mA, maybe a Fast type, but I have no info to confirm.  It could be 1A.

[amptramp]

The flyback boost converter that you have here is good for a general purpose high voltage supply but there are a lot of different topologies with different noise characteristics.

For low current, a charge pump is often seen as the way to get decent performance where you only need a few milliamps and the voltage change is not that much.  There is a lot of noise since turn on and turn off of the switching device happens at maximum current.

Flyback is more efficient and switching on is done at zero current but the sudden switching off is at maximum current and this generates noise.

There is the forward converter that reduces the amount of iron needed since the output current is delivered at the same time as the input current.

There are push-pull converters of various types that can be better for noise.

There is even the current-fed converter that has poor regulation performance but it has one characteristic that is valuable in one application: if all transistors are turned on briefly due to a nuclear event, this converter keeps working because nuclear photocurrents are limited by the inductive feed and the noise is low due to the triangular current waveform which is quieter than any square wave current input.

There are resonant converters that act like sine wave transmitters where the noise output is concentrated at the resonant frequency and it is easy to design filtering for a single frequency, but some of these are also variable repetition rate where a sine wave of a single frequency is produced with some variable dead time between each sine-wave pulse.

Some converters have a fixed frequency and variable pulse width whereas others have a fixed pulse width and variable on time.  It is easier to design filtering for a fixed frequency converter.  You should keep in mind that simple converter designs often need complicated filtering to make the overall noise level acceptable.  The flyback converter proposed here is probably good but will need an input filter that would probably be larger than the converter module and the converter should be designed for a frequency well above the audio range.  If you need more than one converter in a pedal chain, the beat frequency between the converters should be either above the audio range or the converters should be synchronized.

[Rob Strand]

Some converters have a fixed frequency and variable pulse width whereas others have a fixed pulse width and variable on time.  It is easier to design filtering for a fixed frequency converter. 

FWIW,  NE555 = constant off-time variable on-time, MC34063 = constant on-time variable off-time.
For the common HV circuits kicking around there's no current limit or current feedback.
Most simple converters run in discontinuous mode so there's two off-times: an off time for when the inductor current runs dry and an off-time for the switch.

[amptramp]

If I was going to build a tube pedal for myself, I might just use a transformer from the 120 VAC supply with a diode bridge and forget about trying to use an inverter for high voltage.  There is still a possibility of noise and if you are building it commercially to sell, you need CSA or UL and in some markets, VDE approval, so the inverter is a better idea in this case.  If the inverter frequency is high enough, filtering should not be that much of a problem although electrolytic capacitors lose effectiveness at frequencies above audio, so we are talking multilayer ceramic for the most predictable filtering.  High frequencies require fast switching times if you want good efficiency and the usual module you can buy off the internet may not be suitable for the task.  Try it and see, but you will need a filter and you must design the filter for low output impedance which means a lot of capacitance and not much inductance.

[Lorok]

Comparing the traced Kingsley SMPS schematics from crane and Rob Strand, there is a difference in the bottom voltage divider resistor coming off pin 5 off the MC34063. crane's schematic uses 2k67, while Rob's uses 3k3. Going by the datasheet's Vout=1.25x(1+R2/R1) formula, this put's crane's version at 320V and Rob's at 259V before being dropped by the 10k resistor. This seems like a big difference to me, I wonder which one is the "correct" part to get the ~250V at the output?

Another difference is the inclusion of a diode to ground at the input. Next to the fuse it looks like an overvoltage protection diode, and if so it should be a zener with a breakdown voltage of about 15-20V? But the Kingsley part looks a lot bigger than these zeners, so I wonder what the real purpose of this diode is.

Maybe somebody has a clue about these questions? Thanks!

[Rob Strand]

Maybe somebody has a clue about these questions?

From what I can see crane doesn't give an output voltage.   crane indicates tuning the 680k to set the output voltage.  In fact you can tune either.  2.67k + 680k is going to produce a higher output voltage than 3.3k + 680k at C1.    I think you get that already.  (From what I can see 2.67k + 680k is going to put out 320V at C1.)

The extra variable is the 10k on the output.  The voltage at C2 depends on how much current is being pulled from the load.  Maybe it turns out 3k3 is a little under 250V and 2k7 is a little over 250V.  The trace early on in the thread is only showing 200V for the supply.    If we have an uncertainty of 200V to 250V then it's hard to know what is correct.   (If we pull 2mA from the supply then that's 20V drop across the 10k, which favours the 3.3k.)

The diode is just reverse polarity protection.  The diode needs to be a reasonable size and the fuse will blow when the supply is reversed.

[printer2]

Just some information on the 40-390V modules. Used one this week to power a 6V6 and 12AX7. The weak spot on these are suppose to be the HV diode. I blew one after drawing 300V/55mA. I replaced the diode with a pair of UF4007's and 350V, 40mA was fine. Went for broke, 390V/49mA, in free air it was about as hot as I would like. The diode was rated for 3A, the two UF4007's 2A, wondering if they used substandard parts?

[Rob Strand]

Just some information on the 40-390V modules. Used one this week to power a 6V6 and 12AX7. The weak spot on these are suppose to be the HV diode. I blew one after drawing 300V/55mA. I replaced the diode with a pair of UF4007's and 350V, 40mA was fine. Went for broke, 390V/49mA, in free air it was about as hot as I would like. The diode was rated for 3A, the two UF4007's 2A, wondering if they used substandard parts?

Do you know what the original part was?

Layout can promote failures as small parasitic inductances can cause HV glitches which exceed the part ratings on a small time scale and eventually cause failures.   These days dodgy parts is also a possibility.

[printer2]

Do you know what the original part was?

Layout can promote failures as small parasitic inductances can cause HV glitches which exceed the part ratings on a small time scale and eventually cause failures.   These days dodgy parts is also a possibility.

A US3MC. I was thinking of the UF4007's getting hot and unbalance with all the other things that could go wrong. I was really surprised when it worked. Reading about counterfeit chips I would not be surprised if the US3MC either did not have the current capacity or is not as quick as it is suppose to be. The one I repaired seems stable, another one I blew and replaced the diode was bouncing around but I did not spend more time on it yet. Here is the replacements in.



Here is a stock image. There is an additional diode in the upper right corner to give a +/- supply. I thought of taking the upper diode and paralleling it with the existing one but will it cause issues as you said? Since I have a number of the UF's I tried it and it seems ok.



I did try something else before. I CA'd a piece of aluminum onto the body of the diode. It did give me a few more watts until it blew.

[Rob Strand]

A US3MC. I was thinking of the UF4007's getting hot and unbalance with all the other things that could go wrong. I was really surprised when it worked. Reading about counterfeit chips I would not be surprised if the US3MC either did not have the current capacity or is not as quick as it is suppose to be. The one I repaired seems stable, another one I blew and replaced the diode was bouncing around but I did not spend more time on it yet. Here is the replacements in.

It's not obvious to me what the cause is.   I can only speculate.   If you build a large number of boards and they all fail then it could be the layout or the diode batch or a design issue.  If a different batch or supplier of diodes fixes it the it's the diode, otherwise perhaps it's the layout or design.

When you have one-off failure it could be anything.  If you replace the diode with different diodes it might look like it's the diode type but it could be the diode and replacing it with the same diode type would also work.

Early on you mentioned it was getting hot.  I'm assuming it's diode that's getting hot.   That would be a significant piece of information.   When you convert low voltage to high voltage with a boost converter the duty is very low and requires a high current pulse.   It's possible the inductance value is too small.   It could even be getting near saturation are bumping the current up.

[crane]

Sorry for late reply - I don't visit this place as often as I used to. But still - here is an explanation/answer to your questions/concerns.
Input diode? Doesn't harm anything - fullfils reverse polarity protection. I think I had it on my actual build although it's missing in the schematic I have. Schematic will work without it.
Feedback voltage dividers? Different designs use different voltages. I had several of these pedals to open and they had different values. You are supposed to calculate your own values for the voltage you need. That is the beauty of it. Tested ~150....400V (unfrotunately the tests were without load) - worked OK.

Comparing the traced Kingsley SMPS schematics from crane and Rob Strand, there is a difference in the bottom voltage divider resistor coming off pin 5 off the MC34063. crane's schematic uses 2k67, while Rob's uses 3k3. Going by the datasheet's Vout=1.25x(1+R2/R1) formula, this put's crane's version at 320V and Rob's at 259V before being dropped by the 10k resistor. This seems like a big difference to me, I wonder which one is the "correct" part to get the ~250V at the output?

Another difference is the inclusion of a diode to ground at the input. Next to the fuse it looks like an overvoltage protection diode, and if so it should be a zener with a breakdown voltage of about 15-20V? But the Kingsley part looks a lot bigger than these zeners, so I wonder what the real purpose of this diode is.

Maybe somebody has a clue about these questions? Thanks!

[printer2]

It's not obvious to me what the cause is.   I can only speculate.   If you build a large number of boards and they all fail then it could be the layout or the diode batch or a design issue.  If a different batch or supplier of diodes fixes it the it's the diode, otherwise perhaps it's the layout or design.

When you have one-off failure it could be anything.  If you replace the diode with different diodes it might look like it's the diode type but it could be the diode and replacing it with the same diode type would also work.

Early on you mentioned it was getting hot.  I'm assuming it's diode that's getting hot.   That would be a significant piece of information.   When you convert low voltage to high voltage with a boost converter the duty is very low and requires a high current pulse.   It's possible the inductance value is too small.   It could even be getting near saturation are bumping the current up.

Yes it could be a lot of reasons why it might have blown, from poor design to bad parts. First heard of the diode blowing from this video.

https://youtu.be/qB8FpGTyVTw?t=755

Think this may be the schematic.



And just for kicks.



Some information here.

https://www.diyaudio.com/community/threads/high-voltage-boost-converter-sacrilege.310526/page-4

[amptramp]

There is an FCC requirement that any piece of equipment that is not battery-operated must pass conducted emissions testing at a cost of thousands of dollars unless all generated frequencies are below 9000 Hz.  Someone on this forum said he made quiet power supplies that operated below 9000 Hz but did not inject any power supply noise into the circuit.  This might be worth it if you intend to produce pedals for sale.  I have forgotten who it was but it was in a post from this year.  I would be interested in finding out what circuit topology he was using so I could determine if it was suitable for specific purposes.

[duck_arse]

FSFX - he used to bait antonis.

https://www.diystompboxes.com/smfforum/index.php?topic=130799.0
https://www.diystompboxes.com/smfforum/index.php?topic=130687.0
https://www.diystompboxes.com/smfforum/index.php?topic=108238.0

[printer2]

I thought I had a 12V 5A laptop supply to use but it seems to be limiting the current. Would never have tried it but at 9.5V in I was still getting 337V out at 16W. I have a few other supplies but none over 2.7A with 12V. I did get 30W out of the module with a fan on, may have to stop by a thrift store to see if I can find a higher current one.