SKYLINER build - inspired by ODS #124 and #060

[Stephen1966]

So we have some good news... after rewiring the cathodes' 1 ohm resistors in series between the cathode and ground, they work. I tested the cathode pins to ground for continuity (all good) and took a cold reading for resistance. There is some inherent resistance in the cathode/ground circuit meaning they don't measure 1 ohm exactly. After running the amp for 10 minutes or so and then testing for resistance with the amp still warm, but switched off and filter caps drained. Each of the test points came out at 1.13 ohms.

[Richard1001]

What resistance does your meter give with the testwires tied together? More often than not, the testleads and conectors themself have some resistance. (Actualy allways, but some meters can be set to zero)

[Stephen1966]

What resistance does your meter give with the testwires tied together? More often than not, the testleads and conectors themself have some resistance. (Actualy allways, but some meters can be set to zero)

Good question Great question
I put the meter in reference mode and hey presto, the resistance of the test leads is 0.13 ohms. A very small variation between the test points:

V7: 1.00 ohm
V8: 1.05 ohm
V9: 1.04 ohm
V10: 1.02 ohm

Lesson here: to measure Ik using reference mode on the meter.

In real terms, that's something like a 5mA difference, more or less - just about what the screen resistor should be pulling. Now, before sending any more data, I should repeat the experiment tomorrow when I'm not half as drunk

[Stephen1966]

I have one more question, if I may...

Is the relationship between the current draw of the screen resistor and cathode of output tubes linear? I.e. as the cathode current increases, does the screen resistor current increase proportionally? For instance, screen current as a relatively fixed percentage of the cathode current?

I can solve this problem through direct observation, but what is the consensus view of it?

If I understand it correctly, it is a linear relationship. A product or a sum of factors of the plate current.

That's rubbish. Here's some measured data backing up the correction I proposed above:

Re: bias: cathode current vs plate current
https://ampgarage.com/forum/viewtopic.p ... 83#p389683

[martin manning]

I put the meter in reference mode and hey presto, the resistance of the test leads is 0.13 ohms. A very small variation between the test points:

V7: 1.00 ohm
V8: 1.05 ohm
V9: 1.04 ohm
V10: 1.02 ohm

Lesson here: to measure Ik using reference mode on the meter.

In real terms, that's something like a 5mA difference, more or less - just about what the screen resistor should be pulling.

I question the accuracy of even a high quality multimeter when measuring resistances in the range of 1Ω, and just accept the resistor value as being 1Ω within its stated tolerance. Fortunately you will be measuring voltage, in a range where your meter should have good accuracy. Note that if one of your 1Ω resistors really is ca. 1.05Ω, that will result in a sensed current error of 5%, not 5mA.

[Stephen1966]

...
I question the accuracy of even a high quality multimeter when measuring resistances in the range of 1Ω, and just accept the resistor value as being 1Ω within its stated tolerance. Fortunately you will be measuring voltage, in a range where your meter should have good accuracy. Note that if one of your 1Ω resistors really is ca. 1.05Ω, that will result in a sensed current error of 5%, not 5mA.

Agreed. A 5% error for the cathode resistor(s) is within normal tolerances and shouldn't throw the readings off by very much. The 5mA is what some say the screen resistor draws. When I was measuring at 1.13 ohms the cathode current was off (decreased) by ~5mA. I don't know if anyone else bothers with this but when using the cathode bias method the general idea seems to be that the reading you get, will be on the conservative side because it includes the screen current as well. I would like to be able to calculate a more realistic reading by deducting the screen current from the target bias. Hence the question if the screen current resistance is in direct proportion to the cathode resistance. If so, the figure of 5mA is only good for a singular bias setting, and 5% (of cathode current) is good across the entire range of bias settings - being in a linear relationship. Everything tells me it is in a linear relationship and it is better to refer to the screen current as a percentage of the cathode current.

Alas, wall voltages are never stable and even the variac will vary from the beginning of a session to the end. Not by much, granted, but those AC variations are amplified in the B+ and can result in measurable differences across the bias procedure (yes, I'm that slow ). Just looking for a rule-of-thumb, then. And to understand it correctly...

[pompeiisneaks]

If you insist on knowing the actual plate current, the measurement you're using won't give you that, (not suer why it matters, the goal is biasing correctly, and if you do the common assumption that about 5mA is the screen current, then it works very well and has been used for a long time... but I digress)

You should use the more dangerous/harder/less used OT measurement method. It is more accurate.

With the amp off, you measure each half of the OT's windings resistance measurement and write that down (note this method only measures the OT halves and can't do per tube if there are say 4 or 6 tubes in a push pull setup. only each half)

Next connect your multimeter in DC Volts mode to the one half of the winding and the other the Output Transformer Center Tap where it connects to B+. Turn on the amp, and with all volumes at 0 measure the VDC. Wait for the amp to warm up and first and then write measurement down. Repeat for the other half of the OT.

Using Ohms law you can calculate the plate current via:

I = V / R So lets make up some fake measurements:

OT Brown = 15ohms
OT Blue = 14.8 ohms
V Brown = 3.4 VDC
V Blue = 3.45 VDC

Current Brown half = 3.40/15 = .225A or 22.6mA
Current Blue half = 3.45 / 14.8 = .223 A or 23.3 mA

Of course this isn't a real amp, but that's the actual current measured in that transformer coming from plate current.

NOTE: you're measuring HALF of the transformer, this is the plate current for ALL tubes on that half of the transformer. For a 2 tube push pull setup, this is the correct value per tube, but say if you have a 100W 6L6GC amp with 2 tubes per half, that measurement is the plate current for two tubes from that half of the transformer. Thus you need to divide the measured current by the number of tubes per half. I.e. a 4 tube setup needs each half divided by 2 to get the average 'per tube' plate current.

Due to dealing with measuring the full B+ instead of just the cathode current of a few volts, where you're measuring 27.6mV = 22.6mA Plate current + 5mA screen current, you are much safer in general. It's not some major risk if you know how to be safe, but still...

Edit: Fixing some of the measurements I have here, because I was explaining this wrong. You measure from the OT CT which is at B+ to the anodes of the tubes which is usually some small amount of voltage drop over the OT, not hundreds like my sample used to show.
~Phil

[Stephen1966]

Thank you Phil, and thank you all, for your lucid descriptions of different biasing techniques. I will try yours Phil and I will try the others as well. My aim here is not to reinvent the wheel - I think we can all agree that cathode biasing is the safest from the voltage point of view, but also from the practical side as well. Because once the chassis is inside the head, there is really no simpler way to check the bias than using cathode bias technique.

My aim at this point is simply to try all the techniques (bar the transformer shunt method which I consider too hazardous) to find verifiable results, to discover the interactions that take place, and to practice the different formulas and acquire a working knowledge. I still think to assume 5mA for screen current has some problems but you (and plenty of others) could well be right to make the assumption. I would really like to know for myself and until the data is aligned to support either a 5mA or x% I would hold off making any assumptions of my own. Testing, should reveal the relationship. I was only asking for an hypothetical understanding of the relationship between screen/cathode and plate currents. It's easier to confirm results, than it is to test all the variables so to recap...

[martin manning]

The plot below was made from the graphs in GE's 6L6GC data sheet. 5% of anode current is a good estimate for screen current in beam tubes. In true pentodes it is higher, (~14%) as shown in the table I posted a few years ago (which you linked above).

[Stephen1966]

The plot below was made from the graphs in GE's 6L6GC data sheet. 5% of anode current is a good estimate for screen current in beam tubes. In true pentodes it is higher, (~14%) as shown in the table I posted a few years ago (which you linked above).

Much obliged, Martin. Thank you.

I want to plot a graph like this, using several of the biasing techniques and comparing their results to see where they tally... and how they diverge. I can see that the idea of a "linear" relationship between screen current and cathode current is not going to be, strictly speaking linear. Tubes don't output in a strictly linear fashion, but perhaps I will see something like a parallel relationship?

This is the best time to test these procedures. I have the amp on my bench and it won't be going anywhere until I have the head built. Once it goes in the head, I'm likely going to be too lazy to pull it back out and experiment with it. The good thing, is that everything about this amp now (thanks to the great counsel I found here) seems to be working as it should, so (theory be damned) electrons don't lie and I'm looking forward to the learning curve it will provide.

A note for anyone joining this conversation later... I might play fast and loose with my descriptions and ideas but I don't take any chances when it comes to safety. You shouldn't either. Trust me, I know what it feels like to have mains voltage coursing through your body. I still bear the scars over forty years later. A reminder to self. Confucius said, "a wise man learns from his mistakes; a wiser man learns from the mistakes of others." Let that be a lesson for you.

[martin manning]

I can see that the idea of a "linear" relationship between screen current and cathode current is not going to be, strictly speaking linear.

If you look between 50% and 70% Pa Max, where you’ll be setting the bias, it’s pretty darn linear.

[FUCHSAUDIO]

Good to know, thanks. I'm just thinking what else I would need to do if I swap the PI plates' wires, maybe swap the plate resistors as well? Hmm. I'm hoping it doesn't come to that. The output section just after the PI is all but equivalent either side of the PI but the grid and cathodes are pretty well baked in to the circuit with half the stage fed by its grid and the other half by its cathode . Is this 50/50 because of the OT windings or the PT? It was Andy Fuchs who mentioned something about reversing the HV leads from the PT to remove oscillation. Is this not a better solution to the problem than re-engineering the LTP?

If the neighbours complain, I just send the wife out. She's in PR so it's just like homework for her

To be clear: My comment about reversing phase was for output transformer plate leads. If you have nfb, and the amp oscillates, you may actually have positive feedback, and the plate leads need to be swapped. Easy test is to lift the NFB wire from the speaker jack and if the amp gets louder, you have things correct and you have negative FB. You may need to look elsewhere if you have instability or problems with your output stage. If the leads are reversed, the amp would get quieter when you lifted the NFB wires and that shows you have positive feedback, which is wrong. I've seen amps work and sound okay, but "bark" or "tweet" when

[Stephen1966]

Good to know, thanks. I'm just thinking what else I would need to do if I swap the PI plates' wires, maybe swap the plate resistors as well? Hmm. I'm hoping it doesn't come to that. The output section just after the PI is all but equivalent either side of the PI but the grid and cathodes are pretty well baked in to the circuit with half the stage fed by its grid and the other half by its cathode . Is this 50/50 because of the OT windings or the PT? It was Andy Fuchs who mentioned something about reversing the HV leads from the PT to remove oscillation. Is this not a better solution to the problem than re-engineering the LTP?

If the neighbours complain, I just send the wife out. She's in PR so it's just like homework for her

To be clear: My comment about reversing phase was for output transformer plate leads. If you have nfb, and the amp oscillates, you may actually have positive feedback, and the plate leads need to be swapped. Easy test is to lift the NFB wire from the speaker jack and if the amp gets louder, you have things correct and you have negative FB. You may need to look elsewhere if you have instability or problems with your output stage. If the leads are reversed, the amp would get quieter when you lifted the NFB wires and that shows you have positive feedback, which is wrong. I've seen amps work and sound okay, but "bark" or "tweet" when

Thank you Andy, I'm happy to stand corrected... Great tip about the volume with and without the NFB, but "bark" and "tweet"... when? We all love a cliffhanger but the suspense is killing me

[talbany]

Generally the procedure for hooking up the GNFB wire is??..Leaving it disconnected during the initial light off..After you get the amp up and running and put out any fires touch the wire to the speaker jack (Like Andy states) the volume should drop. If it's backwards and squeals? swap the plate leads (leaving enough wire to do so) and solder the wire to the jack and re-bias the amp and your good to go. This generally avoids the startle factor catching you off guard knocking something over trying to turn the amp off or damaging a speaker scaring the wife,dog kids etc etc..(It's usually the last wire I solder)

A little tip I learned from Emmit Clarks Repair shop

Tony

[Stephen1966]

Generally the procedure for hooking up the GNFB wire is??..Leaving it disconnected during the initial light off..After you get the amp up and running and put out any fires touch the wire to the speaker jack (Like Andy states) the volume should drop. If it's backwards and squeals? swap the plate leads (leaving enough wire to do so) and solder the wire to the jack and re-bias the amp and your good to go. This generally avoids the startle factor catching you off guard knocking something over trying to turn the amp off or damaging a speaker scaring the wife,dog kids etc etc..(It's usually the last wire I solder)

A little tip I learned from Emmit Clarks Repair shop

Tony

I think we are generally good to go at this end, if I understand it well though there may be a situation where the positive feedback is not audible and so I can double check this (lift the NFB and then reattach, with a signal) if you think it's wise. Emmit Clark's Repair Shop, sounds a hoot. Knocking stuff over, damaging speakers scaring wives and kids - is this the voice of experience, I wonder. You say it's the last wire you solder though - I liked rootz's suggestion earlier, of leaving the OT secondaries long in case they need to be reversed. Perhaps a good procedure might be to try touching the NFB as you suggest, and then when that's working as it should, finish off the secondaries, tidying them up and finalising them in place. Hey ho! There isn't much that can go wrong that can't be fixed... after all the small fires are out, that is