Simplistic MosFET HV Shunt Regs

diyAudio Chief Moderator
Joined 2002
Paid Member
Yes well, in the Netherlands it is all flat and no rocks and stones. All cables are in the ground except some of the very high tension lines. Not so much fluctuation in consumption, easier to maintain.
And, extremely important: The grid has not been privatised (yet)!
Yes its an easy place for laying infrastructure. I have been to Utrecht, Amsterdam, Rotterdam, and to a southern small town named Hulst only accessible directly within Holland by ferryboat. Was interesting to climb on the sea dam near there and see up-close what "Netherlands" really means. Also interesting how easy it was crossing the Belgian border to visit the close enough port city of Antwerp by just waving hi to the guard from locals he recognized.
 
Oh, you’ve been here? Then you must have noticed: no street poles with wobbling transformers hanging in them like in so many other countries, everything underground.

It must have been some time ago when you were here; since 2003 we have a 6,6 km long tunnel to connect Zuid Beveland to Zeeuws Vlaanderen called the ‘Westerscheldetunnel’. The ferry is long gone. In these days we at least had some border patrol men waving at you, nowadays there is no border at all. ;)
 
Last edited:
Salas

Sorry if i have missed it in the past threads. Up to about page 150.

I have a 1hz noise on my phono preamp. I believe it to be coming from my mains voltage variations.
It appears as not a sine wave. More of just variations at approx 1Hz and varying amplitude. I already have a pi filter of 220uf 10h 220uf. 75ma curret draw. I have even tried 1.4 farads across the pi filter to no avail.

The issue manifests itself as speaker excursion at slightly loud MM phono levels. My goal is to get it to MC. Curcuit is allen wright SVP.

Any advice will be appreciated.
 
diyAudio Chief Moderator
Joined 2002
Paid Member
There could be theories when without experimental data from a spectrum analyzer. Some 1/F thermal noise (near DC) originating in the reg and in the tubes themselves so they add up or its in the tubes only. Depends on tubes samples also not just types.

If its only in the reg you could try add an RC filter cell of few unimportant volts drop on the SVP's rail and see if the phenomenon tends to diminish. That would tell its just sensitivity to the reg's near DC noise and for some zero PSRR circuitry reason the phono gain magnifies it.

Can modify it to best values and keep the RC cell if it does not mess with your sound. If it does and you can't find an optimum coupling capacitors value either that sounds right then you may decide tampering with the reg. But a phono should have a form of subsonic filtering response anyway due to arm & cart resonance, eccentric records wobbling, turntable rumble, etc.

First try isolate its Vref JFET better from thermal shifts by encasing it in some compound. Replace a couple K117GRs first if you got spares so to see if there is a better behaving sample to start with.

The brute force way is filtering with a higher value Vref filter capacitor (now 10uF). Say 22uF or 47uF. That one is rather dangerous as it can kickback higher energy and hit active parts in the reg harder. At least firstly the D3 Zener should be replaced with a heavy (3A) Schottky diode in same orientation to can absorb on/off transitions better.

No guarantee that the reg will be working as trouble free as before though. It may or it may not.
 
I obviously made a mistake. Let me ask you this: I require 380V and a load current of 20 mA. What, in your opinion, should be the right settings?
I may only guess based on a similar design.
I'd consider +/-5% of mains and not more than 2V of p-p pulses after the first filter (the transformer and filter you should model, for example, in PSUD). I'd also taking into account that the SSHV2 will work with around min 10V across the CCS (difference between in and out) and around min 10mA in the shunt section. Yes, as pointed by Salas, there will be some degradation in the output impedance, but it will happen only while there is low mains voltage. So, min input-output voltage should be not less than 11V (I'd use 15V) and I'd take min shunt voltage =15mA. You use a schematic that stabilizes the current, it will help with the possible degradation.
So, max input voltage will be ((380+15)/0.95)*1.05=437V, CCS power: (437-380)*(20+15)=2W, shunt power: 380*0.015=5.7W.
I'd not count the faulty cases. If you really bother about such cases - there are simple circuits that can disconnect the regulator (like "electronic fuses").
 
I have a 1hz noise on my phono preamp. I believe it to be coming from my mains voltage variations.
It appears as not a sine wave. More of just variations at approx 1Hz and varying amplitude. I already have a pi filter of 220uf 10h 220uf. 75ma curret draw. I have even tried 1.4 farads across the pi filter to no avail.
Check (and possibly put a damper) the LC in your filter, the very low frequency oscillations can be originated from here.
 
Well, in this case the SSHV is providing both properties by design.
Such big capacitance is impossible without electrolytics (and this is the first problem with "speed" as far as even best electrolytics have limited frequency response), and even electrolytics in such high voltage are big in sizes, increasing leads' inductivity (another "speed" problem). "Energy"? Yes, it is important, but we have a lot of energy in mains, the problem is a quick delivery to the circuit, which is already mentioned in "speed". On the other hand, the SSHV provides necessary energy by default (CCS part) and the shunt part redistributes the energy with max possible speed between the circuit and the damping section.
 
Last edited:
Well, in this case the SSHV is providing both properties by design.
Such big capacitance is impossible without electrolytics (and this is the first problem with "speed" as far as even best electrolytics have limited frequency response), and even electrolytics in such high voltage are big in sizes, increasing leads' inductivity (another "speed" problem). "Energy"? Yes, it is important, but we have a lot of energy in mains, the problem is a quick delivery to the circuit, which is already mentioned in "speed". On the other hand, the SSHV provides necessary energy by default (CCS part) and the shunt part redistributes the energy with max possible speed between the circuit and the damping section.

Do you think by-passing the big cap with a carefully calculated cap with the best suited porperty (material) can help with the speed??
 
diyAudio Chief Moderator
Joined 2002
Paid Member
Is there a reason why an opamp can't outperform the discreet components in the salas design?

No reason. Surely it could. SSHV is not about super performance, its about a different approach. Here is passive unity gain Norton Vref (no multiplication extra noise) and few select components, intending to simplicity and medium gain wide bandwidth OLG.

Could have designed typically like internal TL783 architecture but I wasn't interested in that.
 
Do you think by-passing the big cap with a carefully calculated cap with the best suited porperty (material) can help with the speed??
Definitely - yes! I should point out that I'm not against Makarov's design (I'll name it this way to make clear what I'm speaking about). There are many ways to achieve the stated goal. I'm speaking about meaningful approach to achieve the destination. You mentioned "carefully calculations". Do you think ba0504 has made them? Analyse the SSHV properties? What is achievable through the addition? From the answers I can say - no!