diytube.com

Hi Folks,

During my rebuild, one of the things I wanted to do was replace that very sucky and dangerous bias arrangement in the ST-70.

Currently, if the wipers on the pots lift, the tubes die. Perhaps some iron too.

We need a method that if any wipers lift, the tubes just simply cutoff.

I also wanted thte ability to use non perfectly matched tubes.

Here's the circuit that so far has proven itself quite useful for doing this:



I've just used regular pots for the controls, one could even use the existing 10K in there for the balance control. But when a PCB is made, I will likely use the multiturn Bourns 3296 trimmers that are often used in pro gear for their reliability.

Cheers!

That'll do it for sure.

Another idea might be add a regulator so the voltage output won't be sensitive to line variations. -50 volt is bit outside the limit of a LM317. Shunt arrangements are best for low noise.

Hi,

davygrvy wrote:Another idea might be add a regulator so the voltage output won't be sensitive to line variations. -50 volt is bit outside the limit of a LM317. Shunt arrangements are best for low noise.

I used a trick like that until a fellow pointed out that you don't want stiff voltage there... if the line voltage goes up or down, you want the bias to do the same and "track" the fluctualtions.

A regulator is fine for pentode mode, where the tube's not as sensitive to anode voltage fluctuations, but in UL or triode mode, things can get scary.

Cheers!

Geek wrote:A regulator is fine for pentode mode, where the tube's not as sensitive to anode voltage fluctuations, but in UL or triode mode, things can get scary.

On a pentode amp, if you regulate the bias voltage but do not regulate the screen voltage, you will have the same kind of problem. The screen supply voltage will wander up and down with the mains. Regulate both, or regulate none.

Ok, yes, you're right, but.. A 10% drop in B+ along with a 10% drop in bias voltage doesn't result in same idle current, but it is closer. A tweaky accurate way is to use the cathode voltage of the output tubes as a feedback point to modify the bias voltage so the current is constant. So instead of a shunt regulator using it's output voltage as the feedback point, the voltage at the cathode is used so it acts as a constant current regulator. Of course, properly slowed to behave at DC and taking into account warm-up. There's probably lots of these circuits in the wild. I've wanted to design one myself.

davygrvy wrote: -50 volt is bit outside the limit of a LM317.

The LM317 can be used for high voltage applications. See the Michael Maida application note. I'm sure there must be some way to get a large negative voltage out of the thing.
www.national.com/ms/LB/LB-47.pdf

davygrvy wrote:A tweaky accurate way is to use the cathode voltage of the output tubes as a feedback point to modify the bias voltage so the current is constant.

You mean like this?

http://geek.scorpiorising.ca/GeeK_ZonE/ ... pic=1728.0

It works!!!!!!

On a pentode amp, if you regulate the bias voltage but do not regulate the screen voltage, you will have the same kind of problem. The screen supply voltage will wander up and down with the mains. Regulate both, or regulate none.

Excellent point!

I've been so long in the habit of regulating everything (many times through an L-C filter, as in my guitar amps) or nothing, it slipped my mind when I made that comment.

Cheers!

Geek wrote:

davygrvy wrote:A tweaky accurate way is to use the cathode voltage of the output tubes as a feedback point to modify the bias voltage so the current is constant.

You mean like this?

http://geek.scorpiorising.ca/GeeK_ZonE/ ... pic=1728.0

It works!!!!!!

Sweet. Needs a mosfet instead of that transistor, though. But, yeah, that's it. Tube swapping for different types, no problem with that circuit.

davygrvy wrote:Needs a mosfet instead of that transistor, though.

Why?

I must try that then

Cheers!

Mainly cause i've been getting my fu with them and that's really about it. Their positive thermal behavior is just wonderful, but not like it has much bearing when operating as a pass device in this case.

/me is loading your circuit in SPICE to see it work.

davygrvy wrote:Mainly cause i've been getting my fu with them and that's really about it. Their positive thermal behavior is just wonderful, but not like it has much bearing when operating as a pass device in this case.

/me is loading your circuit in SPICE to see it work.

I should revisit this one.

I designed it for protecting a trioded EL36 (near vertical curves!).

My experience with MOSFETS have been mixed... despite the hype, they'll go into thermal runaway just like a bipolar


What SPICE you using? I'd love the model if it's CM student.

Cheers!

LTspice IV, free @ http://www.linear.com/designtools/software/ltspice.jsp

Thanks!

I'll load it up on the wife's box (she has XP, I use *nix)

Cheers!

I used a trick like that until a fellow pointed out that you don't want stiff voltage there... if the line voltage goes up or down, you want the bias to do the same and "track" the fluctualtions.

In my ST-70 days I retrofitted at lest two ST-70s with CC Chong cap boards. Those boards used a Zener diode to regulate the bias voltage to -56 volts. If what you are saying is right then there are a lot of people with CC Chong modded units with problems

-50 volt is bit outside the limit of a LM317

Use a TL783 it is good for 700ma at up to 125 volts.