Hi,
I built a power MOSFET-based variable voltage circuit (VVR) for an 18 watt home-made tube amp I built. This allows one to crank the amp so it's "working", without being incredibly loud. This is a known-working circuit, and I have a couple of questions regarding some trouble I've experienced with it.

In use, after a few hours of operation I started to pick up an annoying but pretty low hum (60Hz). Not too bad, but maybe symptomatic of something not quite right.

Then, months later, while tweaking components, I shifted the voltage up and down a few times pretty fast - maybe not good to roll the pot quickly with these voltages? My tube rectifier arced and the fuse blew. Found a blob of solder under the preamp board the fell out, so maybe I shorted something when soldering. Anyway, removing the whole VVR board showed the 12V zener diode was blown. I have spares so plan to simply re-do the whole thing as I've torn it apart. The rectifier survived the arcing.

I am using an SK3675, which is replacement for the NTE 2377 MOSFET that I couldn't get. The pre-filter coming from the tube rectifier (EZ-81) is shown on some schemas and not others....I suspect if might have been the source of the trouble. It shows up on Si rectifier circuits mostly, and I have a tube recto. I thought I'd use it to try to get some additional smoothing before the VVR, but maybe a bad idea?

What I'm looking for is sort of a small explanation of how this circuit operates (like a variable resistance?), what the zener does, why it blew (due to a transient short farther "down the line" that I knocked out after the failure?), and should I eliminate the pre-filter components when I rebuild it (the 220K thru the 22uF cap)?

The input voltage to the MOSFET circuit is in the neighborhood of 310V ripply DC from the tube rect., and output varies down to as low as about 100V. After the output, the adjusted DC heads to the filter caps in the amp as normal.

Thanks for any insight into this question; this is a great forum where you're not treated like an idiot if you don't immediately grasp the subtleties of how things like this work!!

Circuit below:

 

The zener is in circuit to keep the FET gate from going higher than 12 volts above the source. The FET will die horribly if the gate goes to more than 20 volts over the source potential.

The FET is being used as a variable resistance. I have seen tubes used like that. It might help if the FET were mounted on a heat sink. I would suppose the B+ current must be low, but the FET still has to dissipate a few watts.

My experience with NTE parts is that they are very costly substitutes. You might look for an International Rectifier equivalent that will be much cheaper.

How big is that last filter cap? It might be that charging back up over and over it led to the failure of the FET.

 

Hi,
I built a power MOSFET-based variable voltage circuit (VVR) for an 18 watt home-made tube amp I built. This allows one to crank the amp so it's "working", without being incredibly loud. This is a known-working circuit, and I have a couple of questions regarding some trouble I've experienced with it.

In use, after a few hours of operation I started to pick up an annoying but pretty low hum (60Hz). Not too bad, but maybe symptomatic of something not quite right.

Sounds to me that perhaps your filter capacitor(s) aren't adequate, or you might have excessive loading as the amp warms up. Double-check your tube biasing.

Then, months later, while tweaking components, I shifted the voltage up and down a few times pretty fast - maybe not good to roll the pot quickly with these voltages? My tube rectifier arced and the fuse blew. Found a blob of solder under the preamp board the fell out, so maybe I shorted something when soldering. Anyway, removing the whole VVR board showed the 12V zener diode was blown. I have spares so plan to simply re-do the whole thing as I've torn it apart. The rectifier survived the arcing.

The 12v Zener is to protect against the gate voltage going higher than it should, (Vgs<=12v) or lower than Vs less the Vf of the Zener. If the Zener blew, that would indicate to me that the gate shorted to the drain, and the MOSFET is toast.

I am using an SK3675, which is replacement for the NTE 2377 MOSFET that I couldn't get. The pre-filter coming from the tube rectifier (EZ-81) is shown on some schemas and not others....I suspect if might have been the source of the trouble. It shows up on Si rectifier circuits mostly, and I have a tube recto. I thought I'd use it to try to get some additional smoothing before the VVR, but maybe a bad idea?

It's not a bad idea. You definitely need the resistors, or the MOSFET will be dissipating all the power.

What I'm looking for is sort of a small explanation of how this circuit operates (like a variable resistance?)

Yes.

, what the zener does, why it blew (due to a transient short farther "down the line" that I knocked out after the failure?), and should I eliminate the pre-filter components when I rebuild it (the 220K thru the 22uF cap)?

Keep them.
Already explained what the Zener does; you need to keep it in there.

The input voltage to the MOSFET circuit is in the neighborhood of 310V ripply DC from the tube rect., and output varies down to as low as about 100V. After the output, the adjusted DC heads to the filter caps in the amp as normal.

You can get rid of most of the ripple by adding a cap from the MOSFET drain to ground.

 

Thanks guys, I needed a logic check! I didn't think the pre-filtering was a bad idea...actually I did have that cap located from drain to ground. Could it have simply been that the short in the preamp toasted the MOSFET? I had it mounted to the amp chassis w/hardware & thermal compound...

The 'FETs I have (2SK3675) are a lower-priced but just as 'beefy' version of the NTE part, so I have 3 more to sacrifice, lol. I probably should try it w/just the resistors and see what happens...

 

Yeah, if you had a short in the preamp, a cap on the drain of the MOSFET would cause the MOSFET to have to handle a lot of current from the cap to the short. Maybe that's not such a good idea after all.

You could split the 220 Ohm resistor into two resistors, and put a cap to ground between them. That way you would get extra filtering, but if you had a short fault in the preamp, the resistor between the cap and the MOSFET would limit the current through the MOSFET.

[eta]
You really don't have to have a MOSFET with a 900v rating. You could use one rated for your B+*1.5 or 2, so roughly 450v to 600v. The 220 Ohm resistor limits maximum current to 310v/220 Ohms = 1.41A, so you really only need one rated for an Id (drain current) of twice that. Overkill won't hurt, but isn't terribly necessary. The main thing will be the power rating of the MOSFET, and how well you've heat sinked it.

This MOSFET is pretty cheap, but still would be overkill for your project:
http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=FDPF15N65-ND

Even cheaper:
http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=IRF830APBF-ND

About a buck each, and would still be overkill:
http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=FDPF5N50T-ND

 

Hey, cool! I like the idea of splitting the bleeder and doing some filtering there. Just in case..... I bought 4 or 5 of the SK-type MOSFETs when I first put this thing together, so they're not in short supply.

Must've been that short that popped the VVR; I've corrected it and checked out the preamp, it is functioning (I believe) fine...it took out my power tubes, too, so I can't hear the audio, LOL! But voltages are normal up to the output. I'll know for sure when the new tubes arrive...thanks for the suggestions!

 

can y ou post pictures of the amp itself, including the wiring?

 

Hi Paulo,
It's just a standard Marshall 18W TMB (treble, Mid, bass mod). The amp worked just great until I put in variable voltage, but also worked for 6 mos. with it. Then I changed some cathode resistors and bypass caps, created a short, and blew my power tubes and the VVR. I am here to get an idea if what I am doing is correct; nothing that small will show in the pics.
Here are a couple of pictures of the chassis insides, before I installed the VVR.
Enjoy!

 

Do you have an electrical schematic for your project? It might be of help to initiate an electrical forensic analysis........

 

Hi Mr. Martinez,

This isn't my amp exactly, but is very close. Mine is known as a "TMB", treble,mid,bass that the original 18W didn't have. This schema is set up for the 'Plexi' version of the amp; the power supply section is typical of that era of Marshall amps, though. The schem I do have is 'proprietary' since I bought the parts in kit form. I am going to go thru it and re-do it so that I can 'open source' the diagram, it being common knowledge how these work.

Seems from replies I've had on a few forums that the pre-filtering really isn't necessary...the 220R before the VVR is to simulate tube rectifier sag, which isn't an issue if using a tube recto!! So I am going to just run from transformer to VVR, then standby switch.
Check out this schem, it's a nice amp