Bias wiggle trem pulsing indicator and heaters?

[winterturtle]

Still playing with my little Lafayette that has the knockoff Princeton trem.
https://www.tiffe.de/roehren/u255r.gif it still sounds great but...

The indicator light pulses with the trem. Seems cool honestly but the heaters are on the same power taps from the transformer so im thinking this is not intentional or good? I looked for princeton videos and don't think I saw theirs pulsing with trem on.

I've also noticed the trem slows down a decent bit and the intensity drops (indicator pulses less deeply also) when I strike a chord energetically. I dont believe that is normal either.

This is wrong - right? Any ideas on where to start looking?

[Williamblake]

You could add the existing indicator light to the shematic. Also test the oscillator tube for shorts, maybe cathode to filament is connected.

[Latole]

IMO power supply circuit or / and power transformer lack of power.

[tubeswell]

What are the voltages like when you go into the 'low volume' situation?

[pdf64]

I suggest to clip a fast responding meter across the heater supply, and try to assess if its voltage really is pulsing with the trem.

It may just be that the pilot light connection is dodgy, and vibration is setting it off.

[PRR]

> pulses with the trem.
> trem slows down ....when I strike a chord energetically.

If the power tubes idle HOT, and the trem is kicking them HOTTER, all power inside the amp can sag.

Have you checked power tube bias?

[winterturtle]

You could add the existing indicator light to the shematic. Also test the oscillator tube for shorts, maybe cathode to filament is connected.

Interesting -- and yes I should make a digital/updated schematic. My unit has some semantic differences, upp'd capacitors, a few factors that would be best to have visible!

I'll be checking for various shorts off the oscillator tube tonight! Thanks!

[winterturtle]

What are the voltages like when you go into the 'low volume' situation?

I'll see if I can grab a plate voltage reading off the oscillator when I can get it to do that - unsure how quickly my meter refreshes. It lasts like half a second when the instrument output is really at a peak. The indicator light stops flashing and becomes somewhat weak -- then recovers and continues to pulse. Happens very quickly but this is how I remember the phenomenon. More to come!

[winterturtle]

I suggest to clip a fast responding meter across the heater supply, and try to assess if its voltage really is pulsing with the trem.

It may just be that the pilot light connection is dodgy, and vibration is setting it off.

Yeah good call -- I'm sure I will see a change between trem off to trem on if in fact it's the voltage supply & if not, like you mentioned then it should reflect stable. I've dealt with slow refresh rate meters for most of my "DC career" -- mostly an automotive background =)

[winterturtle]

> pulses with the trem.
> trem slows down ....when I strike a chord energetically.

If the power tubes idle HOT, and the trem is kicking them HOTTER, all power inside the amp can sag.

Have you checked power tube bias?

Oh yeah that's kinda scary. Based on the data sheet & measurements I'm getting, I should be idling at about 70% now & have a solid -23V on the control grids as the schematic calls out.

I tried previously to see if I got a more positive control grid voltage reading while it was pulsing but wasn't able to really tell. I'll check and see if one of my meters has a MAX function that could capture the peak value. I have one nice Ideal clamp meter & the rest are digital thrifty-specials =)

Come to think of it I have an old school voltage and dwell meter that may assist me too actually... Leads hooked up backwards I can at least see how the needle is wanting to swing with the trem speed on fully low.

[astronomicum]

Have you checked power tube bias?

Based on the data sheet & measurements I'm getting, I should be idling at about 70% now & have a solid -23V on the control grids as the schematic calls out.

In the previous discussion about fixing the bias, I think you missed the step to adjust the bias voltage after you fixed the bias circuit. You need to check the bias so that you know you are. Just having the bias volatge where the schematic says is not enough. Based on what I have seen with these Lafayatte/Univox units, the schematic voltage, -23V in this case, is most likely too low and will need to be raised (my guess would be to about -26V or so). Go to Robrobinette.com's Tube Bias Calc, find out where the bias needs to be, measure the voltage across the 10ohm cathode resistor, then calculate the current (don't forget to divide the current by two since the resistor is shared by two tubes). You would then need to reduce the 4.7K bias resistor to raise your voltage if needed. You can figure out which value is most likely best mathematically, or you can use trial and error. Clip or tack-solder different resistor values parallel with the 4.7K, starting with about 47K or so, then recheck. Once you are where you need to be, then you can solder the correct value into the circuit board.

[pdf64]

If the HT supply to the screen grids is > 305V, then even with bogey valves, to keep anode current in check, the bias fed to the control grids will need to be proportionately >-23V. 

A lot of people make that same error with vintage Fender schematics.

The VDC across the 10ohm cathode resistor is useful here. Check the 10ohm value too!

Notice that the voltages along the HT supply each have lower values in brackets alongside? I wonder if those are the nominal readings at full output, with the unbracketed values being at idle?

[winterturtle]

Have you checked power tube bias?

Based on the data sheet & measurements I'm getting, I should be idling at about 70% now & have a solid -23V on the control grids as the schematic calls out.

In the previous discussion about fixing the bias, I think you missed the step to adjust the bias voltage after you fixed the bias circuit. You need to check the bias so that you know you are. Just having the bias volatge where the schematic says is not enough. Based on what I have seen with these Lafayatte/Univox units, the schematic voltage, -23V in this case, is most likely too low and will need to be raised (my guess would be to about -26V or so). Go to Robrobinette.com's Tube Bias Calc, find out where the bias needs to be, measure the voltage across the 10ohm cathode resistor, then calculate the current (don't forget to divide the current by two since the resistor is shared by two tubes). You would then need to reduce the 4.7K bias resistor to raise your voltage if needed. You can figure out which value is most likely best mathematically, or you can use trial and error. Clip or tack-solder different resistor values parallel with the 4.7K, starting with about 47K or so, then recheck. Once you are where you need to be, then you can solder the correct value into the circuit board.

So NGL I am no master here but I believe I have checked my bias properly. On the 6973 data sheet - fixed bias - 350v plate voltage (mine read at 335/333vdc) it says my max output is 20w. Since my tubes aren't in the drop down I had thought this was the appropriate process.

So I used that and the calculator you mentioned and I had remembered 70% but its a bit below. 66.5% Seemingly in an appropriate range. I did also desilder the cathode resistor and got 10.9ohm so have been using that 'actual' value. I reentered the numbers and believe I added a screenshot haha.

I had bought a few different resistors and a 10k bias pot but did this math the first time around and was satisfied it was sufficient for now at least. Naturally LMK if I am in error its entirely possible!

[pdf64]

I’m confused, what anode dissipation limit are you using for your valves?
https://tubedata.altanatubes.com.br/sheets/049/6/6973.pdf

[astronomicum]

Notice that the voltages along the HT supply each have lower values in brackets alongside? I wonder if those are the nominal readings at full output, with the unbracketed values being at idle?

Yes.

[astronomicum]

Since my tubes aren't in the drop down I had thought this was the appropriate process.

So it is not on the list, good grief

Can't use the calculator so you need to go by the data sheet.

     Plate dissipation is 12W Max per tube
     Zero Signal Plate Current for Typical AB Fixed Bias Operation at 350V is 58mA for two tubes, therefore 20.3W (350V x 0.058A)
     for two tubes at about 85% dissipation) (20.3W / 24W)

You are at 0.85V across the 10.9 ohm resistor, therefore 78mA (0.85V / 10.9 ohm) and at your voltage of 335V, 26.1W for two tubes (335V x 0.078A), so way too hot.

If you want to be at 70% dissipation, you need to be at about 8.4W per tube (70% of 12W)
At 335V, your current needs to be about 25mA per tube (8.4W / 335V), therefore 0.55V across your 10.9 ohm cathode resistor (0.05A x 10.9 ohm) for two tubes.

[PRR]

....On the 6973 data sheet -.... it says my max output is 20w.

We don't want "output". That is the heat in the speaker.

We want "Max Plate Dissipation".

Which is 12, not 20, watts for 6973.

[winterturtle]

Oh boy have I stepped in it.

So to get more negative voltage through to bring the -23 down to the -26 range I had thought I would want to put the 10k bias pot I have on hand 8n series with the 47k resistor to ground. My idea being that if that value is increased - the bias circuit remains more negative voltage and achieves the result?

The 4.7k shown in the schematic is currently a 6k (looks factory?) And perhaps I can or should remove that and replace it with the 10k pot. I could preset to 4.7k and see where that gets me bias wise - the thought being a lower value there would allow a more negative bias voltage through.

I think I am understanding this part correctly?

[astronomicum]

A 10K is series with the 47K would probably not give enough range and a 10K pot by itself in place of the 6K would not be recommended without protection in case of wiper failure. Remind me what your Vdc is before your 6K and what is the actual resistance of the 6K. Oh, and maybe I missed it but did you replace your output tubes?

[winterturtle]

A 10K is series with the 47K would probably not give enough range and a 10K pot by itself in place of the 6K would not be recommended without protection in case of wiper failure. Remind me what your Vdc is before your 6K and what is the actual resistance of the 6K. Oh, and maybe I missed it but did you replace your output tubes?

Yes a "new" set of used tubes in my case. Hopefully I don't find myself on further dodgy territory here but I am running a much more affordable set of 6CZ5 tubes. In reviewing how I got here -- decided to check the data sheet & the RCA original info is stating a plate dissipation # of 10w. There's a lot of debate threads but I have found a few references that get into the 10 VS 12 - with the 6CZ5 sheet presenting a 'design center value' and the 6973 stating a real maximum value. The good news for me in this particular swap is the 6CZ5 tubes I got were about $10 per so if I toast them then shame on me -- but at least I'm learning lessons as I stumble! I can definitely confirm my socket wiring supports them.

Since finding the backwards capacitor, I actually hadn't measured the voltage after the diode / before the (schematic listed) 4.7k that in my case is a 6k that measures closer to 7k lol. I get -28.9vdc with all tubes removed and in full on position.

In the last 10m I swapped the 6k for a 4.7k and now I am getting -28vdc after that new resistor (~30vdc before)... Nearly -28vdc at the tube socket.

With 12w in the RobR calculator, 361vdc plate-to-cathode, 2 tubes, .735vdc across the 10.9ohm cathode resistor --- looks like I am now sitting t 95.8% and still a good bit too high right? I had been thinking 70% was the appropriate level - but have been wrong before!!!

Adding on-
On a quick fooling session, it certainly sounds happier (quieter) haha -- same or similar pulses in the indicator. I took a moment and cleaned up the indicator bulb and contacts, also no change. The trem also slows & the pulse gets weaker when I hit a chord with gusto. =/

I also did some reading back and am confirming I am still a bit too hot (ie 70% probaly the right goal.) Think I'm close enough now to try and get the pot wired in series with the 47k? I maybe probably have checked the 47k's value so will verify that, probably replace since I have some on hand -- then if the original was in fact 47k or close add the pot. With the interrelation of parts and kinda a web of a system I am not super confident in my incremental tweaks yet.

[PRR]

98.6%, you may as well run resistor bias.

Resistor cathode bias is safer, semi-self-correcting. The tubes cook but consistently.

'Fixed' grid bias lets you idle cool (most of the night you are a lot closer to idle than max) yet get the big power when demanded. Also you can vary current up and down to get tremolo without going (too much) into excess current.

There's no magic %. But surveys of tube amp designs in many wall-powered applications (battery is different) suggest 65%-75%.

[winterturtle]

98.6%, you may as well run resistor bias.

Resistor cathode bias is safer, semi-self-correcting. The tubes cook but consistently.

'Fixed' grid bias lets you idle cool (most of the night you are a lot closer to idle than max) yet get the big power when demanded. Also you can vary current up and down to get tremolo without going (too much) into excess current.

There's no magic %. But surveys of tube amp designs in many wall-powered applications (battery is different) suggest 65%-75%.

Ultimately -- are you suggesting I up the value of the cathode resistor to bring my issue to conclusion? I think I understand what you're saying in a general sense -- but I don't understand the idea of resistor bias.

I am guessing my best options here are to continue to reduce my 4.7K resistor (the supply resistor right after the half-wave rectifier diode for bias current) -- or to increase the value of the 47K resistor that's connected to ground & bleeding off some of the bias voltage -- or to increase the value of the big 2w 10ohm (that reads 10.9 ohm) cathode resistor.

It's like being back in grade school -- gotta hit that 65-75% somehow!

[winterturtle]

You could add the existing indicator light to the shematic. Also test the oscillator tube for shorts, maybe cathode to filament is connected.

Updated drawing with the indicator and some assorted other items that have been updated already.

https://imgur.com/GZYd47e

Pins 4, 5, and 9 on the 12AX7 for the term are connected which seems normal. No other pins have continuity to the filaments =/

[winterturtle]

Did some tidying up and jumped back in for a bit. I swapped the bias circuit 6-7k resistor that was put in place of the drawing's 4.7k -- back to a new 4.7k. I also removed the extra 10k resistors that had been put added (a weird retrofit between the control grids and the bias supply rather than the stock wires.) I'm now getting a higher bias voltage. -26.9vdc on each tube. With:

12W tubes
353vdc plate to cathode
2 tubes feeding cathode resistor
0.693vdc across cathode resistor
10.9ohm cathode resistor
---> I am thinking ~88% dissipation per tube so still pretty hot.
Likewise I still have the pulsing trem issue. Now that she sounds better in general -- I can hear an unpleasant tick with each pulse as well. No bueno!

B1 = 360vdc
B2 = 328vdc
B3 = 272vdc
B4 = 248vdc

Right off the transformer I have 301VAC -- I'm imagining 119.9VAC instead of 115VAC may be the reason for things being a little 'hot' in general?

Still looking for a solution to better measure the pulse I am seeing in the indicator. I did not previously realize that's AC going to the heaters so my mechanical meter can't manage that =/ The plate voltage on the trem tube is all over. ~145vdc at idle (trem switch on but the intensity all down), pulses of course with intensity up,  then really high with the footswitch open / off.

I'm going to reread this thread a couple times and try to make a plan on how to keep trudging forward!