Building EL-84 push/pull amp out of Zenith Radio Parts

[Chuck H]

Using a 20/1 ratio (100k series w/5k dropper) I'd go with the 8 ohm tap. In fact I did with my own amp (which is about a 20/1 NFB ratio). But your PI uses a 10k tail and a 5k shunt. Mine uses a 47k tail and a 1k shunt. Not sure if that makes a difference, but of course everything does. Anyway, I liked my results pretty well. That doesn't mean it's best for you. You can always experiment with changing the ratio or the source after you're up and running.

[DrGonz78]

So I left this on the back burner for a bit but I am back in action. I finally got around to powering up the beast. Not a bad name for the amp... hmmm... "The Beast." I am going with the working name of the amp as "Scurvy Shyster Bastard" as it relates well with my screen name.

Anyway I have some preliminary voltage checks done and thought it would be good to post some of the results. Attached are the working voltages with all power and preamp tubes installed. Measuring 9.2vdc on the cathodes of EL84's.


One thing to note is that my B+ voltage before the zener dropping strip is 326vdc and after it is 290vdc. So I plan on removing a few of the zeners to bring up the B+ a bit. I measured 29.7ma on one EL84 and 32.3ma on the other. That puts dissipation around 70% or a tiny bit above. When I increase B+ I expect that will lower my bias reading a bit too. Any thoughts on this matter would be greatly appreciated.

Main thing is that there is no smoke or any other major problems, just got to start tweaking up the circuit some. I am very thankful for all the help/advice on this build and without it I would have been lost at times. Now it's time to fine tune the power supply and thereafter get the scope out. Once all that is sorted out I will record some guitar to post online too.

[Chuck H]

You should be fine as bias goes. The cathode bias circuit is self adjusting to some degree. All theory and speculation is over now. The voltage is what it is, so... You can use an alligator clip lead to just bypass the one, more or all the zeners and take some voltage and bias readings. You may end up with favorable voltage with NO zeners. Within that voltage range I don't expect the bias to go too hot. Not for cathode bias. 70% is too cool for cathode bias anyway. I know that a functional bias isn't technically "wrong", but I do know how el84's sound and behave. If you end up with no zeners you may wish to change some other power supply resistors for the preamp voltages. Maybe not.

[DrGonz78]

So I went through this tonight and did some quick idle voltage checks while jumping between zeners. I have attached a pic to show the outcome. What I have found is that with 2 dropping zeners it produces the voltages that I had planned out in my design phase of this amp. However, with 330v on the plate I get a bias reading of 37.8ma on the strongest tube and 33.7ma on the other. Note that 12w / 330v = 36.3ma as my max dissipation(100%) for the tube bias. So I don't think that is going to work well. Probably better to have the plate voltage around 307 using 5 zeners in the string and the max dissipation would be 39ma. These tubes would draw 34.8ma(87%) and 31.1ma(80%). What do you guys think about where I am headed on that idea? What percentage of max current dissipation is ideal for the bias? Also, note that I am going to order some more 6bq5, EL84, and even some Russian 6P14P tubes to get different matches.

[Dave H]

When you do the calculations don't forget to subtract the cathode voltage and check the voltage at the plate as there will be a few volts dropped across the OT. I think it will sound great at 330V. Mine is running B+ 336V. Plate 333V, Cathode 11V, Plate current 35mA which works out at 11.27W

[DrGonz78]

When you do the calculations don't forget to subtract the cathode voltage and check the voltage at the plate as there will be a few volts dropped across the OT. I think it will sound great at 330V. Mine is running B+ 336V. Plate 333V, Cathode 11V, Plate current 35mA which works out at 11.27W

Yeah I need to keep at the analysis for a bit on this one before deciding the end result. I did some checks tonight where I measured the voltages from cathodes to the plates. In one case where I had 4 zeners connected I measured 315v at the first filter cap and 314v at the plates. Then I measured 305v from cathodes to the plates so the voltage on the cathode resistor would have been 9v or so. I think you are right and I will spend more time getting down to exact numbers, thanks.

[Chuck H]

Always use the "working voltage" reading from the cathode to the plate (or simply subtract the cathode voltage from the plate voltage). That is your adjusted voltage. For the working current always subtract the screen draw. Measure the voltage across the screen grid resistors and divide that voltage by the resistance. This is the only way to figure for ell84's because the numbers (voltage, current, watts) are smaller than for big bottles so there's a disproportionate difference. That's why there are posts here where people are concerned when their factory amp has the gross dissipation for el84's at 12 to 14 watts. I'd say to set up your plate voltage for ideal (330-ish read to chassis). Then do your bias calculations using the adjusted voltage and the adjusted current. You'll probably be at 85% to 90%. Which is good for that voltage with cathode bias. If you bias to 60% or 70% you may get more crossover distortion than you need to suffer with.

[DrGonz78]

Okay just to update. I got the plate voltage at 319v and that was measured from cathode to plate, currently using 3 zeners on the B+. The bias settings are right around 90% plate dissipation or just a tad higher. I have jammed the amp twice so far and it sounds like a Scurvy Shyster Bastard alright! The first time jamming the amp it was just nice to realize that I wired up everything correctly and it did not explode. My thoughts about the first sound check was that the tone was really nice and the gain was insane! However, the touch sensitivity of the amp was something that seemed just to be unforgiving. It was like the feeling that every little mistake or blemish was coming through, which makes you only sound good when you play something perfect.

So immediately I got thinking how to reduce gain and first thought of increasing resistance in the direct signal path. Still my thought seemed too direct and I sat on the idea for a bit. I was reading something here one night where Enzo was talking about plate/cathode resistors in regards to gain. Obviously that is the whole premise of plate and cathode resistors, setting bias to control the gain of the tube stages. Enzo had mentioned something on the lines that you don't always have to change something directly in the signal path to make a change to the signal. So a light goes off over my head... The first 12ax7 in my preamp I had 100k plate resistor for the first stage and 220k for the second stage. I decided to change the 220k to a 100k after learning more about how that would effect the gain. Just that same week I had come in possession of a 5751 tube and learned all about that tube through websites & data sheets. At this point I wanted to test the amp out to hear the changes. It was like night and day!! The touch sensitivity problems were all gone and it was a fun night of jamming indeed. The amp is a screamer but very musical and not muddy. That 5751 tube in the first tube stage is now part of my amps design.

I still have some things to look at such as what Chuck was recommending on the phase inverter section and even some tweaking of the feedback resistor, etc. I apologize that I still do not have some audio samples to share. Hopefully in the next week I can get on that one and be able to share some sample recordings.

[The Dude]

Sounds like it's coming along nicely. Good on you Dr.! Thanks for the update. I can't wait to hear the clips.

[DrGonz78]

I finally got some recordings of the amp up on a youtube channel. The audio quality is pretty much sub-par but oh well for better or verse. This first video the feedback of the Roman guitar was hard to overcome in such a tight space. This Roman guitar naturally distorts on the 8th fret and has this gainy type tone to it, pretty cool. I did my best to edit the audio and reduce the bad feedback. The last song on this take is nearly feeding back continuously, so take it with a grain of salt. Enjoy!!
https://youtu.be/uoGn9zM2BRM

Here I am with some more... Shows a blues tone going to heavy metal to light gainy clean tone.
https://youtu.be/ZaETYue1_JE

And one more for good measure... Just some nice jazz on this one.
https://youtu.be/zAEjp8lPhII

[Chuck H]

Sounds great. The record mic is compressing things a bit (I can tell the difference ) But it comes through. Good saturation without losing attack and it still has nice ring and fullness on the clean tones. You mentioned 'No reverb, chorus or anything. Just this amp." I don't know exactly what it is (though I've tried to define it, believe me), but some amps will do that for you. Even played clean they just get right under the note and grab it. It's an almost animated effect. EL84's seem especially good at it. Probably because they typically exhibit more intermodulation distortion than other power tubes and that distortion just sounds like an enhancement to our ears. This effect seems sensitive to voltage and you nailed that. Kudos. The amp sounds like music.

[DrGonz78]

Hey Chuck thanks for the feedback and all the help along the way too! Yeah the camera was a Flip HD device and the audio was pretty bad. I kept getting tons spikes on the left channel of the wave form and had to do some major surgery to the audio just to get it to a usable state. I will probably do some more audio examples of the amp and record it through my studio computer, which will sound way better. Now time to build my first head cabinet to house this Bastard!!

Edit: BTW the no reverb, chorus or anything comment was mentioned by Dan Griffin. He is a great Jazz guitarist that was so kind to offer his excellent playing in my video.

[The Dude]

Dr G,

That thing sounds fantastic! It seems very responsive- like you can go from fairly clean to balls out dirt with only the volume knob on the guitar. I'm a huge fan of EL84 amps and I'm sure I would love having a go at your creation. Great sound, great build quality, and very neat and tidy construction. The Scurvy Shyster Bastard is a success. Congrats!

[DRH1958]

Dr G,

That thing sounds fantastic! It seems very responsive- like you can go from fairly clean to balls out dirt with only the volume knob on the guitar. I'm a huge fan of EL84 amps and I'm sure I would love having a go at your creation. Great sound, great build quality, and very neat and tidy construction. The Scurvy Shyster Bastard is a success. Congrats!

Agreed! I'm glad to have seen the final result and the hear the clips. I sort of forgot about this project as it's been a while since it got started but a great end result. Nice build, Gonz!

[DrGonz78]

Thanks for all the feedback ya guys. Also thanks for all the great advice on the way building the thing. At this point I am already plotting to design an amp with push pull with two 6L6 or EL34 outputs. Another amp idea is that I have a few 6550 tubes laying around and want to build one with single ended output. I will also post my final schematic in a few days once I double check it all and post test point voltages too.