vvr ex-bandmaster install

[burt_toast]

woohoo! been meaning to do this for a while. not an amp tech, like a lot of you, so i was a little apprehensive about this. picked up the 12v zener and nte2973 locally - found the rest at wk/hm. made a copper heatsink (works so far), with white heat-goo between mosfet/mica, and mica/heatsink.

my ex-bandmaster is cathode-biased w/el34s - regulating the entire b+ - works amazingly well - no noise, except a tiny bit of dc crackle in preamp vol pot (haven't done the 1M/.1uf trick yet - plan to tho).

i actually love the sound of this amp with the pre-pi mv, and pretty much did the vvr to extend tube life/run cooler when i'm just fartin' around at home - will likely crank vvr and use mv at gigs. i'm a practical guy...

so - question: is there a spot in the sweep of the vvr pot where the mosfet is coolest? all the way up? all the way down? i've heard turning it down only slightly is hottest - thoughts?

thanks again for all the tube-wisdom

scott

[kagliostro]

question: is there a spot in the sweep of the vvr pot where the mosfet is coolest?

May be I've read something about that, but I don't remember where  

something write by KOC  , really I'm not able to remember

K

[HotBluePlates]

so - question: is there a spot in the sweep of the vvr pot where the mosfet is coolest? all the way up? all the way down?

If I recall the VVR circuit correctly, it is a series regulator for the B+. Whatever current is drawn by the portion of the circuit controleld by the VVR always goes through the MOSFET. As you reduce the output B+, then the extra voltage is dropped across the MOSFET.

Therefore, the MOSFET is coolest when it is dropping 0 volts. It will be hottest when dropping maximum voltage.

Another way to look at it is the amp as a whole draws essentially the same power(*) regardless of the VVR setting. But at low output voltage, the output tubes run cool, so the VVR MOSFET is taking the remaining heat.




* Really, as B+ voltage is dropped, current drawn should reduce. But you get the general idea.

[burt_toast]

thx for the reply Kagliostro and HBP,

yeah, i think i've got some old brain cells hanging on to the resistive 'power-soak' notion of: a lot of attenuation=warm, a little attenuation=hot.

...anyway, jeez it works great. can't believe i waited this long to do it. 

thanks again,
scott

[Willabe]

Therefore, the MOSFET is coolest when it is dropping 0 volts. It will be hottest when dropping maximum voltage.

I seem to remember that Kevin O'Conner said the highest dissipation for the mosfet was when the Power Scaling/VVR was set at half?

That stuck with me because I couldn't understand why.

Could it be that at that setting it's dropping 1/2 the B+ and is still having to pass 1/2 the current to the power tubes?

IIRC PRR said it was at 1/2 setting too.



                     Brad     

[burt_toast]

yes, that's where i read that.
thx willabe.

[HotBluePlates]

Let's assume they're right that dissipation is max at half-power. Without doing math to verify, I'm thinking it is because:

If voltage (output power) is near max, there is maximum current through the regulator, but voltage drop across the regulator is minimum. Voltage * Current = Low, because voltage is low.

If voltage (output power) is near minimum, there is maximum voltage across the regulator, but current draw is well reduced because the voltage across the tubes are minimum, and they draw minimum current. Voltage * current = Low, because current is low, even though voltage across regulator is maximum.

If voltage (output power) is medium, there is a middle-value of voltage across the regulator, and a middle-value of current through the regulator. The tubes draw more current because the voltage across them is greater than the minimum-voltage (minimum output power) case. More-than-minimum voltage times more-than-minimum current = maximum dissipation in the regulator.

[PRR]

> It will be hottest when dropping maximum voltage.
> .....the amp as a whole draws essentially the same power(*) regardless of the VVR setting.

No. If the VVR goes to -zero- volts, the tube current will go essentially to zero.

(You may try this on any handy open amp. Put 100K 2W across the standby switch. B+ will drop to 20V-10V. Measure tube current-- it has dropped from 50-150mA down to 4mA, nearly zero.)

(Most VVRs "stop" the B+ before zero. The tone below 10%-5% of full voltage may be very bad, nobody needs _that_ little power, it is best not to have nasty-zones on a knob, so it may not turn-down past 30V or so. That's still so small it does not change the analysis.)

At zero current there is zero dissipation in either amp or VVR FET.

Worst-case must be in-the-middle.

If the amplifier acted like a pure resistor, the worst-case is clear: half voltage.

Tubes run more as 3/2 law. Half voltage *tends* to be one-thrird (0.35) of current. However this is complicated by bias (cathode resistor or voltage-bias). K-resistor tends to make the tube more like a resistor. FIXED bias will make the tube cut-off very quickly if G2 is dropped while G1 stays the same. Practical "fix" bias VVRs shift the G1 bias "somewhat like" the change of G2 voltage, which may be more like 3/2 law, or not, depending on details. (It would be reasonable to change G1 less than G2, leave the tube relatively richer at low power where tube heat is not an issue.)

> turning it down only slightly is hottest - thoughts?

The worst-case is very broad. We don't have to worry about the slight curvature. Set the VVR output at half-voltage. If this is a 400V max amp, dial it to 200V. It should play at the 10W level. Since this is a fix-bias amp, after verifying a happy idle, BEAT it to force 15W-20W fuzzztone out of the honest 10 Watts. Then check MOSFET temperature.

BEWARE!! The MOSFET is at 400V and 200V. Fatal voltages!! In my younger days, I'd rig a rapid discharge and watch the voltmeter fall down below 50V, then stick my finger on the hot-part. If you do this wrong, you get a serious shock. These days $20-$50 buys a perfectly fine IR non-contact thermometer, a much better plan.

It is good to touch the *grounded* heatsink for temperature. However if you didn't get the mounting just-right (warp in the heatsink, dirt in the white goop) the sink can be cool while the MOSFET is cooking.

[burt_toast]

wow, you guys are good.

i know it's apples & oranges, but when i was in high school i made a "power soak" style attenuator with a bunch of sand-filled resistors - an electrical engineer friend told me watch out for heat at the "turn-it-down-a-little" setting - increased current. again, a vvr isn't the same thing, but i think i got that stuck in my head.

on my steel-chassis amp, the mosfet heat was my only niggling concern. i made my own heatsink, first a cute little one out of copper, then a beefier thick aluminum one w/bigger fins. i think it's okay. i was very careful with mounting and white goo.

thanks PRR et al,
scott