HV regulation.
 

I'm looking at knocking up a HV regulator for a valve preamp, not for voltage regulation per se', but for ripple suppression, which as this PSU will provide power for a phono preamp, needs to have very good ripple suppression of over 90dB if possible. The PSU also needs SC protection but not be overly complicated. I can't throw lots of capacitance at this problem, tied a 820u cap/RC filter, it takes too long to charge the cap.Another reason for wanting to regulate is that the tfmrI intend to use looks nice, IE shiny brass but is less than ideal. EG full wave rectified DC V + 470v, falls to 350v loaded @ 10mA, sec is rated @ 60mA.

I've never built a HV PSU from scratch but looking at various HV PSU regs as used on existing commercial and DIY preamps, there seems to be several approach's.

The simplest is in pic one, which is very simple and uses a mosfet and another Q for current limiting.

Another approach is to use a regulator IC like a 317 with elevated ground, a similar approach is to elevate a series pass element using a string of zeners - pic 2, which to me looks dodgy. A string of over 10 zeners seems to me to be problematic, from a temp co POV and possible failure of one zener in the string, not to mention the space numerous zeners would take up on a PCB.

The last approach I've seen is to use a regulator IC, like the 723 with an external series pass element using a HV trannie like one of the BU***, the few circuits I've seen use two tfmr windings though to power the reg Ic and look overly complex. Pic a PSU from an audio research preamp doesn't use a low voltage rail though, but hope you get the idea.

So what to do, some of my thinking might be a bit muddy, any ideas welcome. Andy.

Lastly is possible I'd like to use what parts I have, I have a small bag of BU208

Re: HV regulation.
 

I may be barking up the wrong tree here, but I think that the BU208 is a bit of a one-trick pony, with parameters such as current gain sacrificed in the quest for high-voltage robustness. In its usual TV line-output topology, it's driven with a hefty blip of base current to drive it rapidly into saturation in its role as an efficient switch, as opposed to linear amplifier. As a series regulator, it would need to be driven with a reasonable amount of base current by a device that as a consequence would be a HV power regulator in its own right. In the past, I've used the BU326A as a linear HV pass element, I believe originally orientated at efficient off-line SMPSUs, which has a happy combination of good current gain and yet decent voltage capability. That approach would probably be seen as old-hat now with loadsa HV power FETs around, aimed for the now-ubiquitous SMPSU world.

Re: HV regulation.
 

Ehm , that's a shame,but makes sense, I was tried driving it with a ZTX605 Darlington at lower voltages before cranking up to wkg V/HV, OP current was really weedy, sorry, don't have figures to hand. Anyway a ZTX605 looks no good at full voltage as it has a Vcb0 of 140v, Vceo 120v.

How about if I configured the 208 as a Darlington or parallel it up? I have a few pulls from TV's of HV LOPT Q's, as you say I may be better off with a mosfet, have a few of those.

Andy.

Re: HV regulation.
 

A LM317 with elevated ground would work, but this IC is rated for maximum 40V between input and output. In a HT stabiliser, it is possible that this may be exceeded (it almost certainly will in current limit).

Better would be a TL783, very similar but rated for 125V in-out.

A string of Zener diodes need not be huge - 100V Zeners exist and are cheap. I would go for something like your first circuit, with the 500k potentiometer replaced by 100k between C1+ and R1, and Zeners between R1 and C1-.

This will give fairly good stability. And you can improve ripple rejection by a few μF across the Zeners. (However, I would raise the 3.3Ω to 120Ω so that current limiting occurs at about 5mA).

Re: HV regulation.
 

Will the Tubecad PS-1 described here be enough for your needs?

https://www.tubecad.com/2008/11/blog0151.htm

You can raise its voltage by using the IXCP10M90S as described in the text.

/Martin

Re: HV regulation.
 

I have a good selection of zeners, so possible to elliminate a big string.

I was going to replace the pot as you say with a fixed resistive divider... " I would raise the 3.3Ω to 120Ω so that current limiting occurs at about 5mA). " but if full load current demand is up to 60mA is this not wrong? The original circuit is supposed to have a I of 100ma, I'm missing something here I think. AFAIK Q2 in the circuit of pic is off under normal operation.... Ah, think I've twigged, the 5mA you mention is at the base of Q2, therefore this is "magnified by it's hfe, sorry, transistors aren't my strong point.

I'll have a play and report back. Andy.

Edit, typed before Martins reply... I had a look at that circuit Martin, but dismissed it because of the use of the IXCY10M45S, but I suppose I could use any HV rated mosfet yes?

Re: HV regulation.
 

Hi Andy

Take a look at Duncans's Mosfet regulator. I have built a couple of these years ago and they are still running. I was expecting things to blow up but the protection diodes seem to do their job very well.

Cheers

Eddie

Re: HV regulation.
 

If it's a valve pre-amp you shouldn't need 100mA! A couple of mA possibly, which is why I suggested 5mA.

Keeping the resistor at 3.3 ohms will give you 100mA-plus before current limiting kicks in. This will give 30W-plus dissipation in the MOSFET, which is likely to make it fail (this is after it has got stinking hot). It's generally good practice to have protection levels comfortably above normal running, but not excessively - who wants to burn their fingers on a transistor when they are looking for that short-circuit?

Re: HV regulation.
 

Thanks Eddie, how you doing!? Good to hear from you and know your still out there alive and well : ) I'll have a butchers at that, just knocking up a few circuits on the bench.

"If it's a valve pre-amp you shouldn't need 100mA! A couple of mA possibly, which is why I suggested 5mA" It won't, I mentioned the 100mA in reference to the 3r3 resistor, which I know isn't fixed but is dependant on load. The preamp/PSU will power a phono stage and line amp, which will only draw 10mA tops, however there will be cathode followers after both of these IP stages to "drive" the connecting cable capacitance which may draw more than that. I havn't got to that bit of the design yet, so as the tfmr sec is 60mA I'm using that as my max capacity, but this isn't fixed in stone.

Andy.

Re: HV regulation.
 

If you're intent on using a bipolar transistor as the series-pass element, consider something like a MJE340/MJE360 - they'll handle a couple of hundred volts with ease.

Alternatively, what about 'keeping it vintage' and using a couple of VR-tubes? 85A2 (85V) 0A2 (150 volts) and 0B2 (108 volts) are available cheaply.

Re: HV regulation.
 

Had a quick play this afternoon and killed a tranny when ramping up the OP current, have to get that bit right by more experimentation. BTW, what sort of formula should I use to calculate Re of Q2 in the first circuit?

I was using a G25N40D mosfet as the series pass element, which luckily survived. I'm not fussed about whether the series pass Q is bipolar or fet, just trying to use what I have rather than buy extra components. I have a few MJE340's, so they are an option.

Andy.

Re: HV regulation.
 

I confess that I don't like using solid-state devices (apart from rectifiers) in power supplies for valve equipment. I see too many blown ones come in for repair. I'm not at all surprised you've already snuffed a transistor out. If they fail short-circuit then that can put a lot of volts onto valves which might not be expecting them.

As G6Tanuki has suggested you can take a valve-based approach. They can fail too but because the supply and the load are operating on more equal terms it seems to happen less often. The noise from VR-tubes takes a little handling but it's perfectly possible to achieve very good performance. The two things I would say are i) that the regulator can draw significant current itself, which might be an issue with a weedy mains transformer, and ii) that depending on the circuit chosen one or more valves might need isolated heater supplies, which again can raise mains transformer issues.

Cheers,

GJ

Re: HV regulation.
 

I have used MPSA42 and PHE13009 in a darlington configuration in a capacitance multiplier (simplistically transistor regulator with no voltage reference and a large (100uF or 200uF) capacitor to the base). I used to use them often when working with higher voltages (even sometimes lower ones). Transistors were from RS or Rapid I think. There are circuits around using modern Fets too.

Although I did develop PSU's for over 500V I never worked with more than about 250V per stage so the driver transistor was Ok. I used the darlington arrangement as HV bipolar transistors all tend to have low gain. I generally used voltage multiplication for the final voltage so could smooth and add voltages for the result but have found capacitance multipliers really do work well. Effective capacitance is base capacitance x transistor arrangement current gain.

Re: HV regulation.
 

2N3585 is another worthy HV transistor to look out for as series regulator if there's not too much dissipation- again, combining its high voltage rating with decent current gain so that its error amp doesn't have to use too much brawn. Comes in a quaint but "proper" metal package (TO66) but, as so often with superannuated and unusual transistors, may be pricey and tricky to track down.

Re: HV regulation.
 

I have one of these Andy. Its from Rainer zur Linde. its a stereo version with a common bridge rect ( not fitted) delivers 50 mA each channel.

I made a few of these up and they work well. I have spares if you want one.

Input on the left. two outputs on the right. In the picture you can half see a couple of resistors across the output, they are to float the heaters above ground.

Joe

Re: HV regulation.
 

ionburn's PHE13009 suggestion sounds like a really useful affordable device- I've used a few MJE340s for low dissipation HV applications but this transistor seems like a good'un. As CFLs fade into history, devices like this might become less common-place. I'm highly sceptical of the datasheet's claimed 80W dissipation from a TO220 package, though- maybe if clamped to a slab of silver the size of the Moon....

Re: HV regulation.
 

Thanks all for the useful suggestions. I've various VR valves but had discounted a valve based approach because of the noise issue associated with VR valves, however as mentioned using valves has the advantage that full HT is delayed.

My idea of using three legged fuses stems from wanting to learn more about them and also having several boxes of things including several HV types pulled out of TV's in the past.

There's another approach which is to make a "ripple gobbler" or capacitive multiplier. I'll look into each and do a bit of testing.

Andy.

Re: HV regulation.
 

If your transformer is rated at 60mA, the DC current limit should be at most 40mA (the max. DC current is a factor sqrt(2) below the max. RMS current).

Re: HV regulation.
 

"If your transformer is rated at 60mA, the DC current limit should be at most 40mA (the max. DC current is a factor sqrt(2) below the max. RMS current)." Thanks, didn't know that. The power tfmr HT winding is really weedy, unloaded and rectified etc gives me 470v, depending on how the RC filters are configured. Any sniff of a load, the tfmr lifts it's skirts and runs off crying, EG HT drops dramatically, IE by 100v or more.

Anyhoo... tried a "ripple gobbler" AKA capacitive multiplier today, and whilst not a regulator it's supposed to nobble ripple, see att 1. Whilst it does reduce ripple as advertised I'm having an issue,see pic 2, I think with VHF oscillation. Looking at the trace in more detail, with the scope cranked up on it's fastest time setting - 2uS/div, four peaks fit in 1 div, I calculated this as 2Mhz ish.

I tidied up the layout, shortened leads and popped a ferrite bead on the lead from the MJE340 emitter to BU208 base, this improved things a tad, but the problem was still there. Also tried a 0.047u across the 820u cap, as well as a 1p and 10r "snubber". I suspected the choke, but there is no oscillation at the junction of L and the two 22u caps, IE first LC filter. However note the trace at the two collectors of the Darlington pair - pic 3, oscillation is on the pos and neg part of the cycle.

Andy.

Re: HV regulation.
 

Try putting a resistor of a couple of hundred ohms - ideally wirewound - between the emitter of the MJE340 and the base of the BU208 (where you've currently got your ferrite bead) - followed by a 1nF cap between the base and emitter of the BU208.

That should reduce the VHF gain sufficiently to stop it squealing, while having essentially zero effect at 50/100Hz.

Re: HV regulation.
 

A small output decoupler ( a few uF) to eliminate tendency to HF "springiness" on the output of the Darlington?

Re: HV regulation.
 

Thanks both, I'll try that and report back.

Andy.

Re: HV regulation.
 

An idea I've only seen applied to KV regulation ( as used on submarine cable PSU) is to use saturation of the core via a second winding to keep the transformer output volts stabilised. The two systems I have in mind fed circa 500mA at a few KV regulated on cables carrying approx five 12 channel groups at 3khz spacing with other circuits. Th idea is that as output volts from the transformer rise, the saturation current is increased to make the transformer less efficient and thus reduce the output volts.

Re: HV regulation.
 

I think that the capacitive multiplier circuit shown in post #19 has the potential to become a pulse generator under certain conditions.

If you look at the circuits on Rod Elliott's website you will see that he has placed a resistor between the base of the driver transistor and ground.

This creates a voltage divider which gives more control over the differential voltage between input and output.

His circuits seem to be designed for much lower voltages so you will have to modify the design to suit your application.

Re: HV regulation.
 

Thanks. This is on the back burner for the moment, but I'll give it a try when I drop back on it.

Andy.

Re: HV regulation.
 

It's called a magnetic amplifier.

Careful design of the core and where windings are located is used to inimise coupling of the AC power into the DC control winding.

Saturable cores are also used to make variable high power inductors. Resonated ith a capacitor it can be tuned onto or away from resonance at the operating frequency to regulate an output.

These can be used to individually regulate each output of a multi-output switch-mode supply. The variable inductors go under the name 'transductors'

David

Re: HV regulation.
 

Finally got round round to building a HV PSU using valves this weekend, see att for schematic. It worked straight off without any mucking about which surprised me. I used what valves I had to hand. The 12E1 is overkill for a 300v 40mA ish supply but I wanted a beefy octal valve, anyroad it works.

Performance isn't too bad, 20mA ripple, which could be further filtered prior to any high gain stage as is normal. Another positive is that it's slow powering up, negatives that you can't bung loads of capacitance on it to surpress ripple.

Some questions, re the two pentodes/b tets G2 max, 100v for the 6BR7, 300v for the 12E1. I presume this is 100v with regard to cathode? That's be 60v ish for the both at present. Are there any things I've overlooked or any mistakes?

Andy.

Re: HV regulation.
 

To wrap this up after building various circuits and mucking about I found the best solution was an CLC filter, EG 100u after the diodes, L then 820u = ripple of less than 1mV P-P. No idea what that would mean in terms of PSU Z.

Out of the various circuits I tried the cap multiplier was disappointing and produced some odd results especially the mosfet version which produced massive ripple with 2nd H.

So though not regulated I think it's the best fit for the job in hand, space etc. Still, it's been an interesting exercise.

Thanks to Molly foor typing this out, Andy.

Re: HV regulation.
 

There is something to be said for the CFE* approach - capacitors are quite cheap these days, it's very simple so there's little to go wrong and you don't need all sorts of exotica like multiple floating heater supplies for different valves etc. One thing to bear in mind though is that 820uF @ 300V stores 37 joules which will give you a pretty nasty belt if you should touch it. It may also take a long time to bleed down after you've switched off.

With a capacitor the psu impedance depends on frequency. If the capacitor doesn't have a large ESR then it will be very low, except at deep bass frequencies, and even down there it can be pretty low. You can estimate it from the formula for the reactance of the last capacitor i.e. 1/(2 x pi x freq x C). So at 20Hz your 820uF would have a reactance of just under 10 ohms (or at least the dimension of ohms - if you draw current at 20Hz you will get a ripple voltage of 10 times the current but, unlike a resistive source, with a capacitative one the voltage and current will be 90 degrees out of phase).

Cheers,

GJ

*Informal military term for Capacitor, Flippin' Enormous (I paraphrase slightly ;D).

Re: HV regulation.
 

"There is something to be said for the CFE* approach" Indeed, I like the simplicity. I always put a bleeder resistor across big caps, it produces a nice blue arc when discharging with my cap discharger.

Thanks for explaining the Z, I thought it was more complicated than that, having to take into account the resistance of the tfmr, L etc.

I've put two VR150/30's as a simple shunt reg to clamp the HT whilst the heaters warm up. Not getting 300v, more like 330v but have to play with the limiting resistor - 10k at present. I'd forgotten how nice these look lighted up, a nice blue ish glow, not orange like the smaller 85A2's.

Andy.

Re: HV regulation.
 

Having issues with the shunt regulator, it's not working as it should. I'm using two VR150/30's in series, so should get about 300v +/- 10v right? I'm getting 340v only 20v ish difference from HT loaded - 356v.

I'll own to some confusion about how to work out the dropping resistor value, I thought it was load current + a bit for the reg valves, which is 5mA - 40mA each. So by that reckoning 300/0.030 = 10k, that's 5mA each valve 20mA load I. Tried this, VR valves struck, glowed, voltage out 340v ish. So tried 5k, same, at one point tried 33k and also tried two R's, one to the bottom VR, one on top, same result.

If in doubt read the manual. I have an STC application report which covers the subject so had a read. It says your basic formula to determine Rs = Vi-Vs/Is+IL ... Vi is Vin, Vs is stabilised voltage, Is is VR current and IL is load current. Did the maths, got 1k6 as the result, used two 3K3's, same %*&!? result - 340v.

Another issue I had was the VR's valves flickering after the thing had been powered for about 10mins but I think this was down to dodgy resistors failing in my circuit.

I remember when I first got these valves, I tried powering them and they worked. I also tried two other VR150's, same result.

Andy.

Re: HV regulation.
 

Time to put this one to bed... After more experimentation I've found getting the dropper resistor that limits VR150 current right takes some doing, but have started to get the hang of it. There's a sweet spot to ensure proper regulation and supplying enough current to the circuit powered or load. Another aspect is supplying enough current to the VR valves so that they glow a nice deep blue, but this obviously effects the final regulated voltage, raising it.

In the end I'm not too bothered about the final regulated voltage in this application ( 10-20v either way makes no real difference to the final gain of the preamp) , the VR valves are really only there to clamp the HT whilst the heaters warm, so that Vmax of the smoothing caps isn't exceeded, and the blue glow is a bonus.

Andy.