Simple EL34 SET; advice?

[audio_moksha]

Hello all,

I am in the processing of building an EL34 SET. I have been reading Morgan Jones' Valve Amp book and this is the first time that I am designing my own SET so I want to make sure I doing things right. I'm using a 6SL7 for the driver stage: B+=300V, Rp=240k, Rk=2k and a 100k pot. These values are from the suggested table in the 6SL7 datasheet. I am looking for 5-10W power and this is what I have for the EL34:

B+=425V Rp=5k (I'm using the Hammond 1628SEA @120ma, 48H). I am considering connecting g3 to the 40% UL tap on the Hammond. So I have drawn my loadline on the plate characteristics curve and have Ia=85mA, Vgk=-32V. Do these values seem ok to you all? Will I get around 5-10W from this amp?

I'm also not sure how to go about designing my power supply. This is what I have so far: Power Transformer is the Hammond 373CZ: 325-0-325 @ 150mA. Its going to be centre tapped full wave rectification with diodes. What I'm not sure about is the CLC filter. I know I want the resonant frequency to be as low as possible and the ripple voltage to be as low as possible too. I don't want it to be too complex so I would like to stick with a simple CLC filter. Could somebody help me out with this? Also how do I calculate the value for my power supply dropper resistor. I need one 300V tap after the 425V taps so what current value do I use to calculate the value of this resistor? Is it the transformer rating value of 150mA? So then will my resistor be 425V - (150mA*R) = 300V? I have a feeling that this may not be the right way to do it...

And one last thing: considering that my output transformer is centre tapped so I still need to physically ground the transformer body?

Any help/advice/suggestions will be appreciated by this first timer

Thanks everyone!

[tubelab.com]

I have built a similar amplifier that has been tested with EL34's and other tubes. The schematic is here:

http://www.tubelab.com/SimpleSE_schematic.htm

And the details are here:

http://www.tubelab.com/SimpleSE.htm

Many people run these tubes at about 350 to 375 volts B+ and use a 3.5 to 4 K ohm load. I have tried this, and I have also tried running a higher B+ voltage with a 5K load. I prefer the sound of the 5K transformer and 430 volts or higher. I have tried several different transformers and I have decided on the 1628SEA's for my amp. I run a slightly higher B+ voltage, 430 to 465 volts depending on whether I use the tube or silicon diode rectifier (I have both). I use a low cost Allied power transformer that puts out 375-0-375 volts.

Connect G2 (not G3) to the Ultralinear tap on the output transformer. Connect G3 directly to the cathode. I get 10 watts with 450 volts of B+ and some cathode feedback. You will definitely get at least 5 watts.

Your bias current may be a bit excessive. The maximum plate dissipation for the EL34 is 25 to 30 watts. Most people run them at around 80% of maximum. You figure the dissipation by multiplying the tube current by the voltage across the tube. You will lose 15 to 20 volts in the output transformer due to the resistance of the wire. You did not say whether you plan to use fixed bias or cathode bias. If you use cathode bias, the voltage drop across the cathode resistor can be subtracted from the B+ to find the voltage across the tube. I measure 36 volts across my cathode resistor but I have more B+, 32 is close for your B+ but every tube is slightly different.

Fixed bias: 425 volts B+ - 20 volts lost in the transformer = 405 volts across the EL34.

Cathode bias: 425 volts B+ - 20 volts lost in the transformer - 32 volts across the cathode resistor = 373 volts across the EL34.

Fixed bias 405 volts times .085 amps = 34.425 watts, the EL34 will be glowing red hot.

Cathode bias 373 volts times .085 amps = 31.705 watts, still too hot.

I am known for pushing tubes way too hard and I run EL34's at 60 to 65 mA (.060 - .065 amps). I have 405 volts across the tubes, which gives just over 26 watts. I have had no issues with JJ, Electro Harmonix, Valve Art or Shuguang tubes. I bought 4 "Winged C" tubes. One was bad on arrival, and one died shortly thereafter. The other 2 are fine. Make sure there is good ventillation if pushed this hard.

Some 6L6GC's (good ones), KT88's, 6550's and 6CA7's can be used in this amplifier by readjusting the bias (or changing the cathode resistor.

The power supply can be designed and modeled using the PSUD power supply design software that can be downloaded here:

http://www.duncanamps.com/psud2/index.html

If you just want to build a common power supply using the transformer that you have chosen, set your first cap to something between 25 and 50 uF with a tube rectifier (5AR4) and up to 100uF with a solid state rectifier (I use 47 uF). Resist the temptation to use a monster cap. This will cause high peak currents and noise in the amp that is hard to kill. I would use a 500 volt Panasonic cap here. The choke should be 4 to 10 Henries with the lowest practical DC resistance and at least a 200mA rating. I use the Triad C-14X (6 henries 200mA available at Allied for $16) in budget amps and the Hammond 193J or 193M for nice amps. I use a 100 uF 500 volt Panasonic electrolytic for the second cap. On high end amps I parallel it with a 100uF 370 volt ASC motor run cap. I get these on Ebay (search for "100 uF 370 VAC ASC Motor Run Capacitor").

To get your 300 volts, you use the actual current drawn by your 6SL7's. This will be about 2 to 4 mA per channel but depends on your particular implementation. I have two ways of figuring this out. The high tech way, is to model the input stage to determine the gain, distortion, and tube current. I use Tubecad, but this is not free software. I don't know if this is justified for a single amplifier design. The low tech way is to assume a tube current of about 3 mA, figure the resistor, put in the closest value, and tweak if necessary. I would use a seperate dropping resistor for each channel, and a 47uF capacitor to ground after each resistor. This improves the channel seperation. So, now use 425V - (2mA*R) = 300V, giving a resistor value of 62.5K ohms. I would try 62K or 75K.

Last, but MOST IMPORTANT, ground the case of the power transformer AND the output transformers, as well as EVERY piece of metal that can be touched by the user (including the volume pot). Make sure that there is continuity to the ground connection on the power cord. This will keep you from getting fried if there is a short in either transformer or elsewhere in the amp (not totally uncommon).

I use these suppliers, there are others:

Allied electronics: www.alliedelec.com

Panasonic Caps: www.digikey.com

Hammond chokes and transformers: www.tubesandmore.com

[audio_moksha]

Tubelab thanks you for that most informative reply. I always wondered how the max plate dissipation played into the amp biasing scheme of things; now I know. Also I am using Cathode bias. I still have a few questions Tubelab and I will be eternally grateful if you would answer them.

1. You said you use a 375-0-375 power transformer. I assume this rating is RMS so wouldn't that put you out with a 530V peak output? Is this right or do you use this to account for some loss in the circuit?

2. I will also use a 450V B+: will this be too high considering that I'm using solid state rectifiers? If I do use 450V B+ will also need a 375-0-375 transformer?

3. How do I obtain a plate current=65mA? Do I have to use a output transformer with 450/65mA=7.5k primary resistance??Doesn't the plate resistor control the max plate current?? I'm confused

4. I have seen many amp designs use a resistor (around 100ohms) in front of the grid. You seem to use one too on the TubelabSE. What is its purpose?

5. The 6SL7 has a max Plate Voltage=300V. I am using this tube in the driver stage with a B+=300V, Rp=200k, Ip=1.25mA, Rk=1.3k. This leaves just about 297.5V on the plate. Is that calling it too close?? I've seen Tom Schlangen's design and he uses a 315V B+???
http://www.tubes.mynetcologne.de/roe...el34set_e.html

Thanks once again Tubelab.

[Tubes4e4]

Hi tubelab,

Originally posted by tubelab.com
and I have also tried running a higher B+ voltage with a 5K load. I prefer the sound of the 5K transformer and 430 volts or higher.

I second this observation, and suggest to try even higher Ra up to 10k for EL34SET, with 400-425V across tube (B+ accordingly higher) and about 55mA Ik.

Tom

[tubelab.com]

1. You said you use a 375-0-375 power transformer. I assume this rating is RMS so wouldn't that put you out with a 530V peak output? Is this right or do you use this to account for some loss in the circuit?

In theory, the peak voltage would be about 530 volts. In practice the first capacitor wil never be charged to the peak value. If the cap were charged to 530 volts the diodes would only conduct for a brief instant at the peak of the sinewave. The load is drawing current from the cap all of the time, and the diodes are only charging the cap for a brief moment near the peak of the sine wave. This "conduction angle" varies depending on the load current, capacitor value, and other circuit losses (ESR of cap, transformer resistance, diode losses). We can use a big transformer with very low resistance wire, a monster cap and diodes to get close to the theoretical (RMS * 1.414) value. This would result in the diodes conducting for only a few degrees of the sine wave cycle. This would cause very high peak currents which makes the transformer, diodes and capacitor heat up and be generally unhappy. These short pulses of high current create high levels of harmonics of the line frequency since the conduction begins to resemble square waves. These harmonics WILL find their way into the signal. This is one of the reasons that people have problems with hum in their amplifiers.

That statement refers to the voltage on the first cap in the CLC filter circuit. The load is connected to the second cap. The choke in between the capacitors has a resistance associated with the wire that it is wound with. This can be from 30 ohms to 250 ohms. The lower resistance chokes are wound with thicker wire which makes them bigger, and therefore more expensive.

We like to use component values that tend to put the real value of the rectified voltage at 1.2 to 1.3 times the RMS voltage MINUS the rectifier losses. A silicon diode has a .7 to 1.0 volt drop which can be neglected. A vacuum tube rectifier has a 20 to 50 volt drop depending on the tube type and amount of current through it. This "variable voltage drop" characteristic of a tube rectifier tends to smooth out those current pulses reducing power supply noise. For this reason many people prefer the sound quality of a tube rectifier over solid state diodes. The tube rectifier will result in a lower B+ value for a given transformer. I have both and I can switch between them. With my 375-0-375 volt (RMS) transformer I get 465 volts of B+ with the silicon diode rectifier, depending on the load current. EL34's biased at 60 mA each + 8 mA in each half of the 12AT7 results in a load current of 136 mA. This is about right for my 150 mA rated transformer.

I have been known to put KT88's in my amp (they are in there now) and crank up the current to 90 or 100 mA each. Yes this is too much current, and the transformer gets HOT. I am planning to use a bigger transformer (more current capability) in the final amp. When I do this the B+ voltage drops to 445 volts.

With a tube rectifier I get 430 volts using EL34's at 60 mA and as low as 395 when running KT88's cranked up.

2. I will also use a 450V B+: will this be too high considering that I'm using solid state rectifiers? If I do use 450V B+ will also need a 375-0-375 transformer?

As stated above you can expect around 465 volts with the 375-0-375 volt transformer. You can use a 360-0-360 volt transformer to get about 445 volts but these are hard to find (the Hammond 274AX can not handle the current and will get too hot). You can use a 350-0-350 volt transformer (Hammond 273BX) and you will get about 425 volts B+.

I vote for the 375-0-375 volt transformer. You can use a choke with a high resistance (the Triad C-14X at 150 ohms) to drop a few volts. The Allied 6K7VG transformer is the SAME transformer as the Hammond 274BX but it is cheaper (I won't tell if you won't). As stated before, I have run these transformers much harder than their ratings without incident. I have used about 20 of the Allied transformers over the years. The only failure was one that got wet durring a hurricane.

3. How do I obtain a plate current=65mA? Do I have to use a output transformer with 450/65mA=7.5k primary resistance??Doesn't the plate resistor control the max plate current?? I'm confused

The dynamic (varying with signal) plate current is controlled by the load impedance (output transformer in this case) that the tube sees. A higher load impedance (5K,7.5K or 10K) will result in less power output, but less distortion. A lower output impedance (3K , 3.6K or 4K)will result in more power (to a point) but higher distortion. The dynamic plate current can be estimated graphically by drawing a load line, or simulated in software (SE Amp CAD). This is independent of the static (DC) plate current which is set by the bias on the tube.

When using cathode bias the static plate current is set by adjusting the value of the cathode resistor. Assuming a B+ voltage of about 450 volts, a cathode resistor of 560 ohms, 5 watts will get you close. That is what I use with EL34's. A higher resistor value will lower the current.

4. I have seen many amp designs use a resistor (around 100ohms) in front of the grid. You seem to use one too on the TubelabSE. What is its purpose?

The resistor is called a grid stopper. It is used to help prevent oscillation. It is a good idea to put one on G1 and G2 of most tubes. The resistor on G2 passes the screen grid current. A 1 watt or larger resistor should be used. I have personally fried a 1/2 watt resistor, so I use 2 watt resistors on G2.

5. The 6SL7 has a max Plate Voltage=300V. I am using this tube in the driver stage with a B+=300V, Rp=200k, Ip=1.25mA, Rk=1.3k. This leaves just about 297.5V on the plate. Is that calling it too close?? I've seen Tom Schlangen's design and he uses a 315V B+???

The maximum plate voltage is 300 volts. There will be some drop across the plate resistor and even with a 315 volt supply you will have far less than that on the plate when the amp is in operation. In fact for lowest distortion adjust your resistor values to get 150 to 200 volts on the plate of the 6SL7.

The design that you refer to uses both halves of the 6SL7 in parallel (2 tubes for a stereo amp) They show a 2.65 mA tube current (1.3 mA per half). This results in a 115 volt drop across the 47K plate resistor puting 200 volts on the plate of the tube, well within the published ratings. The paralleled sections results in a lower noise (not really important in this application) and higher drive current capabilities. If you decide to go this route just copy the resistor values from this schematic.

If you want to use 1 6SL7 tube in the amp (1/2 for each channel) you will need a cathode resistor of about 1.3 K as you suggested, but a plate resistor of 100K to get a plate voltage near 200 volts. 100K at 1.25mA drops 125 volts, If you have a 300 volt supply you should get a plate voltage of 175 volts.

I second this observation, and suggest to try even higher Ra up to 10k for EL34SET, with 400-425V across tube (B+ accordingly higher) and about 55mA Ik.

I tried 10K, but I am power hungry so I settled on 5K in the amplifier I am using with my big speakers, 96db 15 inch OB driver(sounds great, very loud, I get complaints from the neighbors). 3K sounded bad (distortion on heavy bass).

I am using a 3K transformer in the smaller amp that I use in my lab with the Yamaha NS10-M studio monitor speakers. I tried both 3K and 5K transformers. The 5K transformers showed better measured distortion AND power output (460 volts B+ with EL34's and 6L6GC's), but the 3K transformers sounded better on the little speakers. I was stumped until I looked at the impedance curves for my speakers. These "8 ohm" speakers only touch 8 ohms at two points in the audio range. They are far above 8 ohms (over 20 ohms) for much of the range. This is an example of tuning the amp to match speakers that are not "tube friendly".

[Eli Duttman]

A way to "dump" unwanted Volts without generating heat in a resistance is to use a 5R4 rectifier. The forward voltage drop in a 5R4 is LARGE, 67 V. at the max allowed 250 mA.

[kstagger]

Tubelab - those SimpleSE boards look excellent - since I'm utterly sick of hardwiring right now (from currently building a complicated linestage) I will be looking forward to the release of your PCBs.

[tubelab.com]

A way to "dump" unwanted Volts without generating heat in a resistance is to use a 5R4 rectifier. The forward voltage drop in a 5R4 is LARGE, 67 V. at the max allowed 250 mA.

Yes this works, and even a 5AR4 will drop enough voltage in this case. The writer has stated that he wanted to use solid state rectifiers though.

Tubelab - those SimpleSE boards look excellent - since I'm utterly sick of hardwiring right now (from currently building a complicated linestage) I will be looking forward to the release of your PCBs.

I have a lot of e-mail requesting these boards. Unfortunately, I just got another one of those "can't miss" deadlines added to my task list at work. Two projects that "must be finished before the years end". I will be working late for the next 3 or 4 weeks, and must make 1 or 2 more weekend trips. I have been working on the manual every evening and while traveling. I hope to have it finished in a week or two.

I get requests for boards early (I tried this with the TubelabSE), but I can't do this. I stated with the TubelabSE that it was never intended to be a beginners project, however most of the purchasers were beginners. Most had no problem building the amp, but I spent about 3 months answering e-mail about what kind of soldering iron (wire, caps, etc) to buy. I answered all of them, and as far as I know all of their amps are now working.

The SimpleSE is intended for beginners, and I have added photographs of EVERY step from the first resistor through 3 different finished amplifiers. Some of this is already on my web site, with more coming soon. The rest will be uploaded as soon as I see the last amplifier (built by a total rookie) and document it. I am told that it is almost finished.

I will start selling boards when I know that I will be available to properly support them.

[audio_moksha]

Hey tubelab,

Let me get this right, so as you say: (B+ - Va)/Rp = Ia and similarly
Ia= Vgk/Rk? The first expression is for the dynamic plate current and the second expression is for the static plate current? While drawing the loadline you plot the dynamic plate current on the y-axes as one end of the load line right?

Also in your previous post you said : Cathode bias: 425 volts B+ - 20 volts lost in the transformer - 32 volts across the cathode resistor = 373 volts across the EL34.

However, isn't the transformer impedance 5k in which case it would drop 5k*Ia which would be much more than 20V?

tubelab you have really helped me understand these ideas. Many thanks to you!

[tubelab.com]

The static plate current is the DC current through the tube with no AC signal flowing (no input, or volume turned all the way down). It could be measured by inserting a current meter is the plate lead of the tube but this is dangerous because of the high voltage. Up until now I have made the assumption that the plate current is equal to the cathode current. This is only true for normal (A1) triodes, but it is close enough for now. The screen grid (G2) does draw a small amount of current, 2 to 10 mA. Assuming a triode, we can measure the voltage across the cathode resistor, and knowing the resistance value, calculate the cathode current. This is the static tube current, also known as the bias current. It is used to figure dissipation. In theory the AVERAGE current doesn't change when signal is applied. However the instantaneous current is changing with the signal. It is this dynamic (changing with the signal) current that gives us amplification. That dynamic current is flowing through the load IMPEDANCE, creating a dynamic voltage drop, and thus by definition power.

You are confusing the plate load IMPEDANCE of the output transformer with its DC RESISTANCE. The impedance is the load reflected to the tube by the transformers multiplication of the (8 ohm) speaker load. In theory a perfect transformer has no DC resistance, and its impedance multiplication ratio is the square of the turns ratio. If a "perfect" transformer had 250 turns of wire on its primary and 10 turns on the secondary the turns ratio would be 25 to 1. The impedance ratio would be 625 to 1. An 8 ohm load would appear to the tube as a 5K load. It would also transfer all of the energy from the primary to the secondary.

Unfortunately transformers are far from perfect, in fact they are by far the most flawed component in a tube amplifier (that is why it is important to get the best ones that you can afford). A real transformer has a DC resistance because of the wires resistance. This is usually from 50 ohms to 500 ohms, and is the source of the DC voltage loss that I spoke of. It is usually the only loss at DC. The impedance ratio is not constant, there is loss of energy due to magnetic and resistive losses, and this loss is not constant VS frequency. There are also hysteresis and saturation effects, but you can dig into this later. For now just assume that a 5K transformer is a 5K impedance to the tube for all audio frequencies, and a small resistive loss at DC. The resistance should be specified for most (decent) output transformers.

I know that there is a LOT of information to learn about vacuum tubes, and there is no way that I or anyone else can convey it all over a forum. I have been building tube amplifiers for over 40 years and I am still learning. Many people assume that this "old stuff" is simple. I have a college level engineering text book from MIT that was written in the late 1930's. It is full of differential equations that I don't even pretend to understand (I have a masters degree in electrical engineering).

Some of what you are missing will come from a better understanding of electronics (load lines and transformers are the same in solid state theory), and some is vacuum tube specific. It seems like a big obstacle now, but it will come if you keep at it. I had built several amplifiers before I understood as much as you know now, but tubes, transformers, and other parts were free at the local trash dump (old TV and stereo sets).

If you are set on building a single ended tube amplifier either of the two circuits mentioned here are good choices. If you are in a hurry, build one of them without deviating too much from the published plans. You will learn much by doing so. Be extremely careful not to fry yourself. We need more students to learn this stuff if it is to continue into the next generation. If you are not in a big hurry keep digging into Morgan Jones. Also see these web sites:

http://www.tubecad.com/ good information about some specific circuits, I use his modeling software to test out circuits before building.

http://boozhoundlabs.com/howto/ more circuits and general info

http://www.members.aol.com/sbench101/ more circuits and truly weird ideas.

http://www.allaboutcircuits.com/vol_1/index.html a good set of web based electronics lessons. Don't ask me why, but vacuum tubes are covered under Volume III - Semiconductors

http://www.aikenamps.com/ A guitar amp specific site. See the TECH INFO pages. Good information on transformers and biasing of vacuum tubes.

http://www.pmillett.com/tecnical_books_online.htm many vacuum tube textbooks on line for downloading, talk about information overload.

I know that there is a good web site that goes into detail about how to draw load lines, but I am at work now, and I can't find it. I must get back to my project now, or I will never get out of here tonight. It is already 6:30 PM here.