Built a Bassman AA864 on my breadboard, to play with the LTP phase inverter (have only built SE and Cathodyne amps up til now), and wanted to hear what two 6L6s would do.... Got it sounding good (and loud!), and was pushing the plate current with the negative bias pot to see what red plating looked like (and I now know what it looks like lol). Funny thing is, when I dial up to 75 mA or so, it starts to "runaway" - the bias voltage starts dropping, making it worse...., and I have to crank the bias up (more negative) to get it under control.... Is this typical of the LTP? Plate B+ is 430V (means 34W/tube @ 80mA), so I know I'm pushing the tubes hard, but that's why I built it - to experiment... All testing is quiescent, into a 100W dummy load. Also, I put the NFB on a cliplead (not sure what side of the OT it needed to go to), but it does nothing regardless of how it's connected.... Thanks for listening to me scratching my head...🙂
Attachments
My Little Runaway
I've had bias problems in the past when I construct a dual bias setup for the output tubes. I adjust one and the other side goes way out of control at some tipping point. Is this a dual bias circuit?
Ccat.
I've had bias problems in the past when I construct a dual bias setup for the output tubes. I adjust one and the other side goes way out of control at some tipping point. Is this a dual bias circuit?
Ccat.
My circuit is single (common) bias - like the Fender circuit. I did see an old post stating a power tube, when pushed to red-plating- will start "off-gassing" and the tube just "runs away"....but I am curious as to how this makes the bias voltage change (going less negative- making the problem worse)? Knowing what the bias current is doing might help explain it. Obviously plate current is rising. If the tube itself increases grid current, and loads down the bias supply, the voltage would go down (less negative) and add to the runaway....?
It could be your power supply design. Most bias supplies can't produce much current. The best solution is to place a fuse on the power supply to prevent a total run away
Red plating, as it is commonly called, makes the Anode equal or near the temperature of the cathode, reducing the transfer of electrons and effectively damping the current dissipation. You are possibly over running the valves making the transfer reduce and giving you odd readings.
Jon, thanks. Reducing current flow should mean a noticeable drop in amp output amplitude? Will this instantly damage the tube, or only if left operating at this level, for how long? Speaking of operation, I've done this measuring and observation with zero input level (quiescent), but couldn't resist inputting a signal, and have seen measured plate current skyrocket (well over 120mv across 1 ohm). Is this why bias is recommended to be set at no more than 75% max dissipation (because averaged wattage when the amp is being played thru has to be higher than quiescent)?
The bias supply isn't designed to run class AB2, so it will only be stiff if no grid current flows. The grid is getting some ion current (from the plate + screen probably) when the tube is red plated, which starts the runaway. It doesn't take much.
Thanks okcrum! Maybe I will put a fuse on my bias supply, or just not push the tubes too far. I plan on putting test jacks on the amp back panel for true bias measurement (on 1 ohms on the cathode legs), so I can set it accurately.... and use matched tubes of course!
Whatever you do, don't fuse the bias supply! If it blows, you leave your output tubes without ANY bias voltage, which is sure to result in total meltdown. If you're getting enough grid current to load down the bias supply significantly, something's off. At the very least, increase the grid stoppers on the output tubes to limit the amount of current they can pull, then investigate the drive issue.
Tubes can get into a thermal runaway state if they get too hot. This is why you see a spec for a Maximum Circuit Value of Grid No.1 Resistance on the tube data sheet. There is usually different values for fixed and cathode bias. If you must run the tubes very hot, you need to reduce the resistance in the grid #1 circuit. Very high power amps use transformer coupling to keep the resistance to a minimum.
Ampkille: you are absolutely correct! If I put a fuse anywhere I think about 250mA on the common cathode ground would be best....and maybe increase the 1500R stoppers on the power tubes. I want to try to make this amp 6L6/6V6 selectable, will test a rectifier tube to see if switching it in to replace the diodes will drop my B+ 40-50V to run 6V6s, and also switch a different resistor in to range the bias voltage as needed.
Loudthud: I'm getting more and more useful information from data sheets, but am not sure what constitutes the resistance in the grid circuit (I gotta feeling the tube itself contributes to it), and does reducing the resistance affect the grid current, or just less voltage on the grid? I'm of the belief that grid voltage controls plate current, and grid current is very small in comparison (unless the tube is pushed too hard?). Thanks all!!!
Loudthud: I'm getting more and more useful information from data sheets, but am not sure what constitutes the resistance in the grid circuit (I gotta feeling the tube itself contributes to it), and does reducing the resistance affect the grid current, or just less voltage on the grid? I'm of the belief that grid voltage controls plate current, and grid current is very small in comparison (unless the tube is pushed too hard?). Thanks all!!!
Been testing my breadboard with 6V6s - goal is to easily switch from 6L6 - and am pleasantly surprised to find the bias is fine when I simply switch tubes! Set for 60mA with 6L6s, changes to 30mA with 6V6s (and no adjustment made). Added a tube rectifier before I sarted swapping tubes, and was disapointed I didn't get much voltage drop (wanted less plate voltage for the 6V6s). With tube rectification I actually have the lowest plate voltage (390) with hot-biased 6L6s, and the highest (420) with mid-biased 6V6s (which I should have guessed would happen - current loads down voltage...). Maybe going back to diodes will keep the voltages up for 6L6s...but either way I am close to calling this ready to place in a chassis... 🙂
- Status
- Not open for further replies.