Converting old tube radio into guitar amplifier with few addons - help needed

The power transformer primary winding has a DCR that should be included
Read and done, see screenshot :)

The AZ11 has about 5V lower anode drop at 20mA than the AZ41, so there is a significant difference
I understand, but I don't feel like defining the AZ41 from scratch. What my engineer mind says though, is with that voltage drop in mind, we can expect even lower output voltage from the AZ41 - correct?

The idle plate-cathode voltage of the EL41 is less than B+ due to cathode bias, and OPT primary resistance. Max design rating for anode and screen is 300V, but that could be exceeded in some equipment for idle conditions.
So, we design the output stage for a voltage value, which supposedly will be there after the output transformer and such, thus the difference vs. actual supply voltage - that's what I understood and of course, I'm going to read on it more soon.
Yet still, if I may ask you for a simple answer (yes/no is great) - are we exceeding the permissible supply voltage for the output stage/other stages (are we safe to power it up with the predicted ~326V)? It's not that I won't go and educate myself, but it feels better to have a confirmation from a proffessional.


Other news, I got mail today and I can continue wiring things inside. Also, I found a better multimeter and I am waiting for it to arrive - it is a True RMS one and has accuracy specified as 0.8%+5, so I hope it will be enough (Uni-T UT89XD). I know, I could always get a better one since I'm buying new, but there's a funny thing to it, that everything up to 4x its price is actually the same thing. But it's still an upgrade and more functionality I haven't had in my current multimeter.
The IEC mains socket is still two weeks away, and since I feel the build is progressing in a steady pace, I might think of a safe workaround to the lack of it, so I would be able to power it up. No rushing, though.

Much thanks for your help!

Cheers,
Jakub
 

Attachments

  • PSUD_CLC_lowload_RT.png
    PSUD_CLC_lowload_RT.png
    56.8 KB · Views: 86
You still need to estimate the idle Vak for the EL41 using your updated PSUD2 B+ level, and the estimated idle current and the voltage drops across the cathode resistor and output transformer primary winding resistance. Only you can fill in all those details. But it looks to me like you are going to be sitting at about 300V. Valves at idle are pretty forgiving - they have to be as mains voltage could easily move idle Vak around by +/-5% (ie. +/-15V).

Testing will always benefit from having more than one meter connected for monitoring levels - I have two cheap ANENG 8009's that I find indispensable for that (although they needed alternative test leads and crocodile clips). Just be aware that the meters can only show lowish frequency ACV, so if there is oscillation going on you may not know about it. Also the meter likely can't measure low DCV when there is high ACV, or vice versa - which can be the situation with valve amps.
 
Great. Thank you. :)

So now I wanted to get into the theory and design, and as for now, my brain effectively got deep fried and I am unable to comprehend a thing. I never thought I'd trip on basic and potentially easy matters.

Baby steps.

So, I just thought I'll post whatever I figured out today, and that is the output transformer parameters with and without the destined speaker. Accuracy as provided by the cheap LCR tester. It uses some combination of current rise in time and time constant measurement.


Speaker: R DC = 2.3Ω, L = 0.05mH

Output transformer:

Primary: R DC = 216.6Ω, L = 14H, secondary: open
Secondary: R DC = 0.5Ω, L = 0.03mH, primary: open
Primary: R DC = 217,9Ω, L = 7.15H, secondary: speaker connected
Primary: R DC = 217.4Ω, L = 82.5mH, secondary: short circuit


Looking at the results, I'm thinking about the transformer ratio and if it isn't something like 2:1, but I can't believe it would be this easy. Or wait, maybe 19.364k:1? (Z of the speaker to the Z of primary with speaker connected, @1kHz) As said, my mind is fried.

Then, the biggest part, where my mind is frying today, is... impedance. We should calculate the output stage using impedance, not resistance. I see everybody talking values, like if the impedance wasn't defined by frequency, as resistance is.
I could calculate the impedances, probably (|Z| for Z = R+jwL), and assume the values everybody are talking about are specified for the usual known-for-me standard frequency of 1kHz. But because nobody mentions that, who am I to know? :confused: Of course, I tried to find anything on this topic and failed (at least today).

Generally I'm looking at The Valve Wizard website (mostly) and some other random sources I can find.
I'll also have to translate the EL41 datasheet to what the DIY guides are talking about, since the designators for all the parameters like to differ here and there. And figure out how to start the design from the OT, not the tube itself. Wish me luck I guess :D
Next time I'll try to post what I've accomplished with the tube power stage design.

All the best,
Jakub
 
Jakub, imho keep searching for and soaking up on-line references to get a broader awareness. Often forums have a thread or two on good 'references', and many forums have threads along the line of query you have about output transformer 'impedance ratings'. Sometimes vintage authors have presented many articles, like Crowhurst, which are aimed at magazine readers, so not as structured as say RDH4.
 

PRR

Member
Joined 2003
Paid Member
Speaker: R DC = 2.3Ω, L = 0.05mH
Secondary: R DC = 0.5Ω, L = 0.03mH, primary: open

Probably not "mH" here. Probably slipped a dot or a zero.

Primary: ...L = 14H, secondary: open
Secondary: ...L = 0.03mH, primary: open

Taking 0.03H, then 14H/0.03H is 467:1 of impedance. Taking 4 Ohm speaker, this is 1.9k to the tube. This is reasonable for the lower-voltage home radios; in my land, 35L6 etc.
 
Hello!


So I took a break to cool down and I am back with ideas - hopefully they are good. :D

I printed some datasheets of the EL41- to have some graphs to draw on and I did not spot I have some great graphs here, until now - see attached. My brain is also (finally) connecting some dots.

I think I need to figure the exact parameters of the OT (output transformer), because the design process is always relying on its impedance (actual or reflected - I think reflected) and I can't just assume "it must be a 10k one" or so.
I forgot the easiest way to figure out a transformer ratio is just simply powering it up. Of course, this isn't a mains transformer, so I'm thinking of grabbing a function generator, which hopefully would be capable of driving it strong enough to get dependable results. And ideally - a full grown-up RLC meter, which would do some calculations on the go for me.
A question to you - should I check it on say, 50Hz, 1kHz or make a full measurement chart? I mean, at the end of the day it's still an OT with an air gap, so the question might sound rhetorical to you. Similar thing is for the speaker, but I remember from seeing some impedance charts, that the speaker rating is (almost?) always given with 1kHz in mind.
I'm still looking for an exact answer, but the voltage approach seems to be good.


Back to my "figuring-it-outs", I found the Valve Wizard website to be the easiest to read and understand for now, so whatever I wrote above is based on its contents and mainly these lines:

Transformer manufacturers generally state that an output transformer will show 'x' primary impedance with 'y' secondary load. For example: 5k primary impedance with an 8 ohms secondary (speaker). This is derived from the voltage and current ratio of the transformer, which are inversely proportional to each other. That is to say; if the voltage is stepped down, the current must step up:
The turns ratio can be given by:
Vin / Vout = Iout / Iin
Since each is the inverse function of the other, squaring the result will give us the impedance ratio:
Z = (Vin / Vout) squared.
> Hence my "connect a voltage to it" idea;

With our HT of 300V, and taking the data sheet value for maximum anode dissipation as 25W;
Zout = 300^2 / 25
= 3600 ohms
> And my attempts to draw something, which currently show, that the load line might be less steep than the datasheet 7k one - and I think it means higher gain; though it's actually unknown without the OT actual (design) impedance.


Still, after even such brief reading, I feel more confident to just recreate the circuit from the factory schematic and.. try? I mean,
The transformer primary has very low DC resistance so only a few volts will be dropped across it, which we usually ignore. Therefore, the quiescent anode voltage will be the SAME as the HT, no matter how much quiescent anode current we choose.
and also:
But it looks to me like you are going to be sitting at about 300V. Valves at idle are pretty forgiving - they have to be as mains voltage could easily move idle Vak around by +/-5% (ie. +/-15V).
therefore I can also say sorry for being a non-believer. :worship: We learn new things everyday.
Well, somebody designed it and we can't say they did it poorly, if the company wanted the radio to last.
I also have the test thingy, which allows me to monitor the current and limit the mains power noticeably (basically a low wattage light bulb in series with the load), then I can also purposedly find an EL41 which wouldn't be a tragic loss in case of some failure and kindly ask the circuit to work (finally). Don't get me wrong though, as I won't if I see some concern from your side. :happy1:

Back to the design, I'm still lacking a power rating for the speaker, which could hint the overall output power, and with the OT impedance found, I'm sure it would be easily calculable for my supply voltage.


(...) this is 1.9k to the tube. This is reasonable for the lower-voltage home radios; in my land, 35L6 etc.
Looking at the Va/Ia graph, at 2k we would exceed the maximum anode dissipation, therefore I think we may consider my previous results unreliable. Still, thank you very much for your input!


I think that's all for now.
Thank you for your replies and I'll be waiting for more - and hopefully also come back with more.

Jakub
 

Attachments

  • EL41_Pout.png
    EL41_Pout.png
    70.6 KB · Views: 64
Last edited:
Read and done whatever I could, some of which was probably unethical to do, but I did it with all the care I had. I mean, it's weekend time now, so I'm trying to use this time by not going to the office to measure things.
Please do correct me if my thinking and/or results below are wrong.

If we can assume a 4Ω resistor is the same as 4Ω of impedance (technically it is, but in audio we have the speaker, which impedance is varying throughout the frequency), then see below.

Side note: I used a 25W incandescent bulb to limit the power and protect the OT in case something went wrong. Cold resistance was 170Ω, the bulb did not glow, so I would assume the resistance was quite constant.
Also, the new multimeter was garbage - all measurements were lower than they should - so I'll trust my old one for now.

For a reminder: Primary DCR = 217Ω, Secondary DCR = 0.4Ω

Voltage applied to primary, then secondary measured.

Voltages:
Case 1. w/bulb: Vin = 236V, Vout = 5.6V
Case 2. directly: Vin = 242V, Vout = 5.7V

With load:
1. 214V / 4.4V
2. 241V / 4.95V

Thus, by the power of
Valve Wizard said:
The turns ratio can be given by:
Vin / Vout = Iout / Iin
Since each is the inverse function of the other, squaring the result will give us the impedance ratio:
Z = (Vin / Vout) squared.
Impedances:
1. 1802.5:1, w/load: 2370:1
2. 1776:1, w/load: 2365.5:1
For ease I'd say it's either 1.8k to 1 or around 2000:1.

The speaker has around 2.2-2.5Ω of DC resistance. It may not be a standard 4Ω rated speaker, but it seems close - maybe it is, maybe it has more at 1kHz, but the DCR of speakers varies and is never the same as the rating. I wish I could measure it.

Datasheet specifies 7k Ra. I have:
a) 4Ω into OT -> 7.2k reflected (1,8k) or 8k;
b) 2.5Ω into OT -> 4.5k or 5k, but normally we wouldn't calculate it for DCR.

My guess is: The design is fairly close to the datasheet circuit.
check EL41 datasheet for class A examples - note how close the cathode resistance in datasheet is to your radio schematic.
I also noticed that! :happy1:


With all that, I'm just thinking, what the cathode resistance will do? The variation I see here is 150-170Ω. The old carbon-composite resistor could even be worse. The range is not big, and - I still have hard time understanding biasing calculation - would move the bias point quite unnoticeably. So if I'm not wrong here, then would there be any difference sound-wise? (looking for the 2nd harmonics)


Last thing would be the screen (grid 2) resistor. This is where the Valve Wizard page looks completely random, to be honest. I calculated a resistor value of 666Ω :)D) for peak anode current (75mA) and dropping 50V on it.
At the same time, the screen on the factory schematic was connected directly to B+. Something tube specific? But I remember,
The screen connection to the CLC filtered B+ is fine, and about the only addition could be a screen stopper (as you've shown) if you overdrive the class A output.
so I guess I'm still adding it.
I currently have no multiple power supply filtering stages, so I thought, can/should I make a separate stage only for the screen grid? Or is just the resistor (grid stopper) needed?

...but not least - with around 325V on B+ we are exceeding the max ratings by the 25V. Is it okay?
Soon I will try to power up the supply section, see how high it goes with minimal load.

~~~~
I think that's all for this post, it's so late here I forgot what to say more.

As always - thank you for your help and replies!
And have a good week. :)

Jakub
 
Some advice. I have converted a number of radios to guitar amps.

Firstly if possible use a new chassis and tube sockets.

Make sure power supply and fuse is safe as has been discussed.

Follow basic champ or valve junior design.

Layout components so they can easily be changed and are not crammed in.

Don’t try for too much gain. Use lossy tone controls or interstate voltage dividers.

Use small coupling caps to avoid woofyness muddy bass.

Don’t try to over design. Get a feel for the circuit by listening to it and playing it.

Change one thing or stage at a time.

Be patient and keep revisions of your design as it evolves with voltages and component values.

If it’s not sounding right replace your tubes, worn tubes may get more worn and you need to ensure the tubes you are working with are not the problem.

My 2 cents worth