High Voltage Regulator

[Birstall]

I have a Linn Sizmik 12.45 subwoofer. It is powered by a Linn Class V 500 watt amp.
After rectification 250 - 300 v DC is spit and delivered to two 500uF 200v capacitors and a VB 408 voltage regulator. The Vout on the schematic for the regulator is 12v.
The voltage regulator has blown and no longer available, Mauser have no alternatives. They are still available on Ebay but repeating this failed design seems foolhardy
This is a common problem with this particular amp and usually solved by installing a modern plate amp at a cost of $500.
I am a novice but I am curious to know if there is a workaround to solve this issue. Obviously 500 watt SM amps are still around so how are they designed with lower voltages
Your comments and advice will be most welcome

[Pulsepowerguy]

I can think of at least one approach to making a discrete version of the VB 408. The part looks like an integration of some fairly standard items. It would take quite a bit of effort - maybe a fun time for a EE hobbyist. All it really does though is take the rectified line voltage and drop it down (with tons of loss).

A pragmatist would take small 15v 'wall wart' supply and use a linear regulator to take it down to 12V. The wall wart should be a chunky, 60hz type (not a switcher) in order to satisfy the audiophile. This would be highly reliable.

[TheMG]

7812 with a suitable resistor in series with the input to bring the voltage down to something it can handle.

If the load is non-constant add a zener diode across the input to clamp the voltage down if the load get very light on the regulator. I would put the zener there regardless as a precaution anyways.

Seen this sort of thing done in plenty of designs where a linear 3-terminal regulator needs to run from a supply voltage greater than it can normally handle. This avoids the use of oddball rare parts like the VB408.

[Pulsepowerguy]

7812 with a suitable resistor in series with the input to bring the voltage down to something it can handle.

If the load is non-constant add a zener diode across the input to clamp the voltage down if the load get very light on the regulator. I would put the zener there regardless as a precaution anyways.

A resistor in series with a 7812? With 250-300V upstream? For a current on the order of 40mA at 300 volts input the power dissipation of a 7k resistor is on the order of 11 Watts. The big problem though is what happens if the input is 250V. For that case the resistor is too large and the regulator is starved. A 50 Volt variation in input voltage makes this approach fail, to say nothing of line surges or dips. At any rate, I would not go that route.

Birstall: Do you have a schematic of the power supply?

[coromonadalix]

I had tl783 regulator in mind,  but they are i think obsolete too,    would an ac-dc adaptor be good ??     an 120-240 vac to 12vdc

[shakalnokturn]

If you want to follow the low efficiency route you could use a capacitive divider rather than resistive one at least that avoids the heat to dissipate.
Still a dedicated transformer sounds a better approach.

[magic]

Capacitive divider doesn't work for DC.

I would prefer the solution with additional 15V transformer and postregulator. It's trivial, it's efficient, it's reliable.

If you want to stay true to the original (i.e. inefficient and expensive), cascode a 7812 with some TO-220 depletion mode MOSFET, SiC JFET or similar species, bootstrapped to the regulator's output.

If you want cheaper and cruder, use a standard enhancement MOSFET instead and tie its gate to a low power shunt regulator at 20V or so.

[shakalnokturn]

I'm aware that a capacitive divider will not work (or only for a very short time) on DC, neither will a transformer AFAIK  , intended use was before the rectification of course.
Totally agree that a small transformer would be the most suited if there is enough room available.

[David Hess]

If you want to stay true to the original (i.e. inefficient and expensive), cascode a 7812 with some TO-220 depletion mode MOSFET, SiC JFET or similar species, bootstrapped to the regulator's output.

That is what I would do but there is no need to use a depletion mode FET although that is the simplest way; a zener diode can set the bias voltage on a conventional bipolar transistor or power MOSFET and I have done this many times.

One thing to watch out for is that during fault conditions, the regulator's full current will be applied to the high voltage transistor so it is advisable to either include this factor into the power transistor's derating or include a separate current limit.  Vbe current limiting is usually sufficient but old designs tend to use foldback current limiting to limit stress on the high voltage transistor.

There is a trick which can sometimes be used with integrated voltage regulators.  If the bias circuit for the high voltage series transistor sets the voltage at the input of the regulator high enough, then the regulator's built in safe-operating-area protection will implement foldback current limiting helping to protect the high voltage transistor.  This assumes the the regulator's current limit is low enough however.

[CaptDon]

You do see that the vb408 is a 'nightlight' or 'standby' supply
device which does nothing during the actual 'run' mode. The
14.5vdc supply will overrun the vb408 when the amp is up
and running. During the redesign, take care to observe what
your kludge circuit will do when the 14.5vdc comes up.

[magic]

You know, DIYing a high voltage downregulator could be moderately tricky, but adding a transformer-based auxiliary 12V PSU is pretty straightforward.

It seems you would need about that much:
- a 12V transformer or AC wall-wart
- rectifying bridge like BR402 (perhaps smaller) or four diodes to make one
- 35V 330µF or larger capacitor
- one 7812 chip
- one more diode to protect the 7812 from backfeeding when the LM2594 converter comes up

Or alternatively, since your amplifier is already rendered impure by the presence of the LM2594:
- a single 12V switching DC wall-wart with a two-prong mains plug (three prongs will be instant smoke)
- a Schottky diode, I doubt that the circuitry downstream would care about seeing 11.6V instead

[Birstall]

Thank you Magic, I appreciate your response. I wonder if you could draw that out for me if and when you have a moment.

[magic]

Hmm, there is actually another question.

If the original regulator was blown, it may have applied full input voltage to the 12V circuits.
What actually is powered by that 12V rail? Do you have a schematic of that? Did you look for signs of damage to those parts?

[Birstall]

On the second drawing in my original post, it shows that rail going to VCC

The only visible sign of damage was a burn mark below the vb408 and a blown electrolytic capacitor, which I replaced. Whether there is additional component damage, I couldn’t say at this time. I was hoping to solve one issue at a time.

If you are implying that I shouldn’t be wasting time on the voltage regulation when there are other issues is probably correct. However, if I understand you correctly, using a transformer may not be so difficult. in which case I would be interested in understanding exactly how to do that but without a drawing I can’t imagine how it all goes together.

[magic]

Forgot about this thread.

It looks like there is a whole bunch of circuitry there and it may be busted if the 12V regulator blew. Was that exploded capacitor on the 12V rail too?

Did you verify with a DMM that there is no short from the 12V rail to ground?

[TheMG]

Can you post the complete schematic of the power supply, or a link to it?

[magic]

Two images in the first post, if you want just the power supply.

[TheMG]

Two images in the first post, if you want just the power supply.

Those only show small portions of the primary side of the power supply circuit.

[winniethepooh_icu]

Do not give up on this VB408. If 25C ambient and 300V input you can not exceed 6mA through this device before risk of thermal shutdown.  Consider that failure may simply be to lack of heat sink. You are not looking at a burn mark from failure in the picture. You are looking at the effects of long term heat.  In the picture the only heat sink is a small copper area on the board.  Improve heat sinking ability with a aluminum heatsink and it should not fail.  Do not short the heat sink to anything, it will be at the same voltage as Vin, maybe put a hazardous voltage label near it as well.

[magic]

Those only show small portions of the primary side of the power supply circuit.

Interesting idea. I assumed that here is no secondary and this is it, the amp simply runs on rectified mains.

[drvtech]

You could use an MP174 non isolated switching regulator. I've found them excellent except when they have absolutely zero load. This is easily solved by putting a small (500mW) zener across the output. But if you always have at least some load (less than a mA) they do the job.

[Birstall]

I was thinking along the same lines it did last for over 12 years.
Unfortunately I was not able to find a similar VR from a reputable US supplier. The alternative was buy from China a VB408

[bdunham7]

I have the schematics but are there copyright issues in posting them on here? I'd be more comfortable sending a PM

I'm pretty sure posting here for the purpose of discussion would be fair use, if US law even applied to an Australian forum.

[Birstall]

Here's what I have

[TheMG]

That's a rather interesting power supply design...

It almost looks like the regulator's sole purpose is to provide power to the primary side switching control circuits until the power supply starts switching, then it gets power from the auxiliary winding of the main transformer, labelled as "BIAS1" and "BIAS2" (strange naming, as it has pretty much nothing to do with biasing of anything), this is rectified and regulated down to 14.5V by a switching buck regulator. Since the 14.5V is greater than the 12V that the VB405 regulator is set to, it basically supplies no current at all once the SMPS is up and running normally.

Since the linear regulator is only supposed to supply power for a brief amount of time during the initial power-up, that would explain the lack of heatsinking.

This also brings about another question: is the buck regulator working as it should and supplying the 14.5V it's supposed to during normal operation of the power supply? If this circuit is not functioning, it would explain why the linear regulator ran so hot and eventually failed, as it would be powering the SMPS control circuits full-time instead of only at startup.

What I would do first, is to apply about 20VDC to the "VBIAS" test point with the unit disconnected from mains, using a bench power supply. Then check for 14.5V VCC. If that's good, check for a switching waveform at the gates of switching FETs. That way you can assess is there is other damage or issues with the primary side of the power supply, in a safe way without ever applying mains voltage to it.