Installing VVR into a fender champ

[gutsofgold]

Has anyone had any experience playing a smaller SE amplifier with a VVR installed? Is it worth it? Or should I just build some more fuzz pedals? I'm debating between installing a VVR that can vary the entire amps B+ vs. varying just the power tube. Not a lot of decent youtube videos showing the kind of distortion you get from this.

[Groovenut]

Put on your ear plugs, send the family out for ice cream, plug in your guitar, turn the volume to 10 and play your champ. That's the kind of distortion you'll get 

Seriously though, on a circuit like the Champ I would vary the entire circuit's B+. Same goes for any non-master volume tube amp, if you want how the amp sounds overdriven loud but at a quieter, more friendly level.

[PRR]

For *Champ* (and Epi Jr) size amps:

Find a 100K *2 Watt* linear pot. This will not be any guitar/pedal part, you go to Mouser or DigiKey or Newark. (Sometimes the same pot has both a 2W military rating and a 5W commercial rating; that's the one you want. Sorry I do not remember the spec-number.)  While there, get a 0.22uFd 400V capacitor.

From screen B+, series-wire a 2K 1/2W, the 100K pot end to end, and tack a 10K from there to ground.

Wire the pot wiper to 6V6/EL84 Screen grid. Wire the 0.22uFd from screen to ground.

Full-up, it runs full power, a mighty 5 Watts output.

Half-way (150V-180V on screen) it runs about 1 Watt output.

You can turn-down to around zero, however the sound may cut-off before the pot hits "0". That's OK but it annoys users. The 10K bottom resistor can be changed so that "0" is very-very-very weak. 5K or 22K may seem better, depending on the specific tube and your need for gnat-size power.

This leaves the 6V6/EL84 plate at 300V-380V, but turns-down the Screen, which turns-down the Current in the tube, while still leaving good signal grid swing. Because the screen grid draws much less power than the plate, you can do this with a simple (albeit hi-power) pot, instead of a full-blown active voltage controller.

If you just had the pot set half-way, the high resistance would let the screen flap with signal and the gain would be way off. You need a bypass cap. 10uFd would work. However some unfinished calcs and tests suggest that an under-size cap works against the OT and tends to a slight bass-lift at lower powers, which lessens the "thin" sound of low volumes.

[gutsofgold]

PRR, thats an interesting way to approach this but I'm not quite sure how one method is better than the other. I thought the whole purpose of the VVR was the decreased voltage swing.

In your method: Lowering the screen voltage on the 6V6 allows you the achieve the same plate voltage swing with a much lower gate voltage swing, which brings earlier clipping and thus distortion.

With conventional VVR: Linearly decrease of the 6V6 plate and screen voltage together achieve the same thing, no? You're gate voltage swing stays the same but now the plate voltage swing has been decreased, resulting in less total swing before clipping.

Are there benefits to keeping the plate voltage the same and simply decreasing screen voltage?

[PRR]

The stage is loaded. If current is restricted, voltage must also be restricted.

This argument does get dubious at speaker resonance, where impedance is large and little current is demanded. My small experiments say this may be "useful" by emphasizing deep bass at low power. However this would depend on the speaker used; a big 45Hz Fs 15" theater speaker would be different from a 180Hz Fs six-inch.

The advantage is simplicity; but only practical for small amps. (A 100W Marshall could need a 25 Watt pot, which gets silly.)

[gutsofgold]

Sorry to revive the old topic but I purchased everything I need aside from the MOSFET. The Dana Hall VVR kit includes NTE2973 which is hard to find (and expensive when you can).

From searching old threads on AX84 and other forums it seems the following have been or can be used: IRF740, IRF820, or IRF830. I am trying to figure out which characteristics make for a good MOSFET to use in this setup. In this case the setup is a basic source follower. From my understanding MOSFETs dissipate a lot of heat while in linear mode (ie not switching). But what if they aren't being used an amplifier and are acting as a buffer? Are they dissipating great heat?

It seems like so long as all operative conditions fall below the limiting of values of the FET, you then have to look at thermal resistance and thermal coefficient. This makes sense as talked about briefly in the following ax84 thread, the NTE2973 has very low thermal resistance where as the cheaper IRF types are often a magnitude higher.

http://www.ax84.com/bbs/dm.php?thread=377952

PRR, or anyone else knowledgeable on the topic... any thoughts? Is it worth trying the three MOSFETs I listed?

[davent]

I've used IRF830's in a couple of small 6aq5 SE amps in conjuction with Kevin O'Connors Power Scaling circuit. So far so good, (as in years). They were 10 for $6-7 when i bought them.

Board with pot.

[gutsofgold]

That's a cool looking build. Is that power scaling circuit similar to this one here?
http://luisamark.com/mark/files/Power%20Scaling%20for%2018%20watt.gif

It seems like the extra NPN transistor is acting as an active load/constant current source. I need to dig into that some more to see if that helps with making the choice of MOSFET any less constraining.

[davent]

Yup, that's it, Mickey with Kevin's okay put that up on 18watt years ago. Used it in my 18watt but also added a sag control as outlined in one of KOC's books.

For the 6aq5 i put another resisitor between the pot and the MOSFET drain to limit the high voltage as the voltage will go as the 6aq5 wants it lower.
dave

[PRR]

> which characteristics make for a good MOSFET to use

You need Voltage and Power.

Even a Champ will run over 350V, so you need a 500V rating. This is critical: other MOSFET parameters degrade when processed for high voltage, so there are a lot of 50V and 100V parts. You need 500V here, and 1,000V for the bigger (higher-Volt) amps.

The MOSFET will dissipate the difference between what the amp sucks full-up and the lower voltage you choose to run at. This is HEAT. As an absolute worst-case, when turned-down to 5%, it could dissipate ALL (95%) the power that the amp runs at. Heatsinks bigger than your big tubes. In fact it won't ever be that bad. For self-biased amps like Champ, the worst-case is slightly less than 1/4 of the power tube's full-up power consumption. If you believe a 6V6 is a 12W Pdiss tube, the MOSFET feels 3W at most. Fender runs some of these amps at 17W Pdiss, so you might say 4+W dissipation in the MOSFET.

While a "12W" 6V6 can really throw 12 Watts, a "20W" MOSFET will NOT throw 20W. First it needs a Heat Sink to throw more than a Watt. The "20W" is on an infinite sink, which we don't have. You probably want to be looking to 40W-60W MOSFETs in any case. But you also need to understand heat sinks, which is a rather complex field.

For more fun: traditional MOSFETs have the Drain on the metal tab on the heatsink. This will put the heatsink up at 350V. You can not bolt the heatsink to the chassis. You can not touch the heatsink or you will be shocked. Insulators are problematic. Today we can find "fully encapsulated" devices with plastic-encased tabs. I sure would look for that, but the MOSFET legs are still hazardous.

Current (often a key MOSFET spec) will not be an issue. The skinniest hi-power MOSFETs around tolerate way more current than any tube amp. Likewise R(on) is a non-issue.

Take Dave's IRF830. First, it is marked "Obsolete"; it works fine, but in bleeding-edge designs there are probably newer choices with better performance. FWIW DigiKey shows zero in stock and last-call is next month. Mouser has 800 in hand at $4 but shows EoL. Odd-lot vendors will have it for years to come, and maybe cheaper by the handful.

NOT plastic encapsulated, so you will be fooling with insulators and "live" tab.

The 500V rating is ample for 300V-400V tube amps; some of the bigger beasts should use a 1,000V part.

Rated 100 Watts, so if you mount it on copper block with cold water through, one device would handle nearly the biggest tube amps (and the exceptions may all be over 500V). For the 4 Watts it may throw in a Champ, I would be comfortable with thick aluminum 2"*2", exposed both sides to not-hot air flow.

The 4.5 Amp rating is 100X what a Champ will pull, 10X what the biggest beasts ever suck, so non-issue.

Switching speed, gate capacitance, diode recovery rate, all total non-issues in this non-switching application. Transconductance will be specified at a sky-high current (2.7A!) far out of our world, and anyway we hardly care (we turn the gate voltage to-taste, not to a specific number).

[PRR]

> extra NPN transistor is acting as an active load/constant current source.

No. It is a current-limiter.

If output current in "R33 2W" causes more than 0.6V voltage drop, NPN turns-on, clamps MOSFET Gate to Source, MOSFET turns-off. Actually it could balance at a steady output current.

However it is not clear -to me- what value "R33 2W" is. If we read it as 0.33 Ohms, the limit current is 1.8 Amps. This is well below the rating of any large MOSFET, good. And the peak turn-on current into the first filter capacitor wants to be "infinite", which would be bad for the MOSFET. However the peak current is really limited by PT winding resistance and is unlikely to approach 2A. Meanwhile *if* something goes wrong around the first filter cap, the MOSFET could be working at high current AND high voltage, i.e. High Power, and melt of its own heat.

Things don't usually go wrong, and when they do a few-$ MOSFET may be the least of the damage.

This seems to be scaled for a fairly large amplifier. For a Champ, it seems to me that "R33" could be a lot closer to 10 Ohms (and "1R" more like 100r 1W) for lesser MOSFET abuse.

V-A Limiting is an old technique which can limit power device abuse. But it needs hard-thinking and also much specific information.

[davent]

That resistors a  0.33 ohm as per TUT5 where he's power scaled an amp running four 6L6's. In TUT6 he gets around to showing a formula to better tailor the value of that current sense resistor.

For the 6aq5 i'm using a voltage doubler on a 120v secondary so low 300's peak and then set the max from the scaler for high 200's on the 6aq5 plate.

The 830's came from Mouser a number of years ago and next time i looked the price was a fair bit more.

dave