EL84 and 6V6 in parallel???

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hello everyone,

I've been working on some simple single ended guitar amp designs over the past few months and I really want to try running a 6V6 and an EL84 in parallel for the power amp section! I've drawn a rough schematic and attached it to the post, so can someone take a look real quick and let me know if I'm on the right track and if this can be done?! I know I'll need an ideal primary impedance of 3000 ohms on the OT, but the hammond 125ese can be had with a 2500 or 5000 primary so that should be close enough right? As I said before the schematic is purely conceptual, I haven't breadboarded anything yet, I just want some people's opinions on whether this can be done with no issues and also if there are any recommended changes to my design.

Thanks for your help and comments in advance!
 

Attachments

  • el84_6v6_power_amp.pdf
    32.8 KB · Views: 965
Member
Joined 2003
Paid Member
This is for a guitar amp where distortion and sonic signature rule. Certainly they would not share the load well but thats the idea right? I would try it. Worse thing that will happen is to have a little magic smoke escape.
The only thing I would do is kick up the value of the bypass caps unless you want them to add to your distortion characteristics and you don't need the gain.
 
Last edited:
Hey,

actually there are commercial geetar amps that put dissimilar tubes in parallel, though I've only seen push pull so far. Mesa Boogie for example, 6L6+EL34. The mismatch in plate resistance between one of the tubes is a sonic feature for guitar, rather than a 'problem'.

So I would suggest just to try it out!

Kenneth
 
Hello everyone,

I've been working on some simple single ended guitar amp designs over the past few months and I really want to try running a 6V6 and an EL84 in parallel for the power amp section!

In your schematic the tube with the lowest impedance will carry nearly all of the load and the other will just be there and look cool but do little else.

You will need to put power resistor in series with each tube where each resistor is about the impedance of the other tube. This will cut the power of the amp in half and wate a lot of heat but then each tube will be sharing the work equally. Also each output tube will need it's own grid resistor
 
In Parallel, both tubes will become a new "effective" tube...
The transformer sees this "effective" tube as a single unit...it does not care ....
These are not BJT's where you have to worry about Vbe and other matching issues... SO you will not get into the whole current sharing or runaway issues you get with PTAT devices...
You may want to try cap to cross couple the cathodes so they are joined in the AC sense but still have DC isolation for independent biasing... This is if you want to use the feedback of the cathode resisitors, if not , then stick with the independent ones you have now...

2K plate load should be close to where you need... Going to high a plate load will cause much earlier square waving and kill the power output and loss of headroom will occur from the load line intersecting the 'triode" region of the curves, when it should be just to the right of the knee of the 0-bias curve, thus staying in the saturation region of the curves, therefore maintaining the plate resistance from significantly droping off...

Best Regards
Chris
 
Wow thank you everyone for the quick responses! I'm glad to see that the general consensus is "go for it":)!

cerrem - can you please explain what you mean by " You may want to try cap to cross couple the cathodes so they are joined in the AC sense but still have DC isolation for independent biasing... This is if you want to use the feedback of the cathode resisitors, if not , then stick with the independent ones you have now... " Although I've been using tube guitar amps for several years now I'm kind of a newbie to tube audio design. What advantages/disadvantages are there to what you suggested doing/my original design in regards to the sound the amp will produce when amplifying my guitar? Is what your suggesting simply having both cathodes share a single resistor and bypass cap?

ChrisA - What values should I use for the grid resistors that you suggested?

SGregory - How much do you think I should increase the values of the bypass caps? I just used what was common in some single ended amps that use each of the tubes (6v6 or el84) in the power section.

Once again thank you for all the replies!
 
When it comes to guitar amps...do whatever you want to make it sound good, so long as it is safe... The classic guitar amps were "monkied" to get working...there were no scientist or engineers doing any hard core analysis or design process...It was all clue-less mishaps and dollar signs in the eyes of these amp companies that just copied from Leo's stuff... 90% of todays amps are all just a mix of copied circuits, copied effects loops, copied master volumes, copied gain stages with little or no engineering behind it..
As far as paralleling tubes... the first order overly simplistic way to see this as two current sources in parallel working into the load... The transconductance of the tube makes it a non-linear voltage dependent current source..
Both tubes contribute current into the load, thus the current from each tube sums to make one contribution into the load...
Looking at this in a more realistic manner, the tubes are not ideal current sources and therefore contribute some some load to the other tube in parallel to the transformer load... So you have a current divider effect...
But this is OK, since the plate loading of one tube in parallel with the transformer load still constitutes a valid load for the other valve to work into..
El84 figure roughly 38k Ohms of plate resisitance and the 6V6 at roughly 50K ohms in your operating range....
Now figure a plate load of 1K ohms ...... Since a 2K plate load is close to optimum fit for each valve if it were independent Single-Ended amp... Since you have two valves in parallel, thus your plate load will half as well..
A 38K ohms current source in parallel with a 50K ohm current source in parallel with a 1K load.... Give me a break, there is not much current divider action going on between the tubes, 99% of the sumed AC current will contribute into the 1K Plate load... This of course refering to mid band performance...

Best Regards
Chris
 
Member
Joined 2003
Paid Member
Youngb4,
20uF will work but values such as 220uF or higher are not uncommon at that value of cathode bias resistor. Experiment to get the sound you want.
This amp sounds fun! At first go I wouldn't worry about feedback too much. Make that version two. It can be rewired and added later if this concept pleases you.
 
Yes i cant wait to start tinkering and experimenting with it! and im still going back and forth with the pre amp! Im thinking ef86 to TMB tone stack to a couple of 12AX7 gain stages, of course i wont know how that sounds till i get all my parts sourced and get a lil more free time on my hands!
 
Yes, Yes, Yes!!!

Incredibly, the idea of combining 6V6 and EL84 has been done before and it sounds terrific. I owned a Mesa Research Tigris for many years, (yes, a part of the Mesa Boogie company that produced audiophile amps), that was the most fun and best sounding amp I ever had. Don't ever let people dissuade you from using your intuition about how different tubes might sound together. You're absolutely right, the tonal balance of 6V6 and EL84 together is wonderful. Think of it more like a musical instrument than as accurate audio reproduction and you will enjoy it much more. For rock this combination is great.

As far as balancing different tubes the Tigris used 2 parallel el84 and 1 6v6 on each side of the OPT. It used standard cathode biasing resistors for biasing with 2 separate resistors, one for the el84 pair and one for the 6v6. Both were bypassed by capacitors. There is really no problem with doing this and it works fine if you get matched tubes. If you don't it is a problem.

If you decide to build this amp I would suggest you do away with cathode biasing so that you can use tubes that are not always fully configured for the amp. Mesa always required you to get tubes from them for that reason. If you put in tubes that don't match current wise for a certain cathode bias it would overheat and blow (don't ask me how I know).

There is one big problem though if you go to direct biasing and that is that the 6v6 and el84 require different bias levels. But it doesn't mean that you need two different driver stages even if you go for grid biasing. What you should do instead is create grid biasing for the el84 (about -11 volts) and then install a 8 volt drop in the 6v6 cathode via leds or zener/led combo.

I plan on building a better version of the tigris for my own listening pleasure because I became addicted to the live sound of music it reproduced. I'm going to drive the final stage with a source follower circuit and use the biasing implementation I just described. I also plan on putting in pots in the cathode circuits with small value resistances (less than 10 ohms) one for each el84 pair and one for each 6v6. To do this circuit right requires some added complication to the original tigris circuit but the sound I got from the original convinced me it is well worth it.

Let me say one more thing. Life teaches you when a lot of people scream that you shouldn't do something it means something very significant. That is never a neutral sign. Usually it means you shouldn't do something (like hurting people). But sometimes it means that you should try it and that most people have been brainwashed about the subject. It is subliminally threatening to them that they have been brainwashed and that is why they react so strongly. I think the second alternative applies here.

:trapper:
 
Member
Joined 2003
Paid Member
This is single ended amp so having a matched pair or quad isn't essential or required. In a PPP or (is it a 1/2PPPP in the Mesa) having matched tubes on each side probably would be important.
Being a single ended, individual "auto bias", as it is originally drawn would also work fine provided the B+ and bias current fall within the capabilities of the tube.

Build it and see.
 
I have built guitar amps with two different output tubes, but they were push pull amps with one 6V6GT and one 6L6GC. Another interesting trick is to replace the grid resistor with a pot on one tube in a P-P amp. This way you can unbalance the drive without changing the bias, allowing for some unusual distortion effects.

I have not tried two dissimilar tubes in a SE amp yet. I am sure that it will work. As stated some experimenting with cap values will be needed to get the sound you want. Generally in a HiFi amp large coupling and cathode bypass capacitor values are desired since you want extended low frequency response. This is not the case with a guitar amp. Too much bass response can overwhelm the speaker and make it sound muddy. This is dependent on the speaker being used, so you need to experiment.

If I were doing this I might be tempted to make one tube dominant for the low frequencies and the other dominant for the treble and add a balance control to adjust the split between them.
 
This is great! I had no idea that this thread would spark so many responses and recommendations - thank you all who have replied, it's going to be a huge help once I sit down and start experimenting!

Tubelab - I love your suggestion about making one tube more dedicated for lows and the other for mid/highs with an adjustable balance control! I have an Egnater Rebel 30 (guitar amp) which uses a pair of el84s and 6v6s that you can blend in between (which is where i got this idea :) ) and I love that feature - so how would I go about doing something like what you suggested? Like I said - I'm a bit of a novice when it comes to some of the design aspects. : /

SGregory - Go for it! If i can play - anyone can! I am a firm believer that a good guitar amp is an extension of the instrument and it's a blast experimenting with the different tonal qualities of each design

Stalker - Thats great! Please let us know what you find out! Did I mention I'm also going to experiment with power scaling the power amp : )
 
I expect you could get something passable by biasing the 6V6s class A, EL84s class C. Because EL84 has more transconductance, they will start conducting at higher signal levels. This will make a sort of crossover distortion while retaining useful gain at small signal levels. The harmonic effect will be to add middle harmonics, i.e. lots of 2nd, 4th and 6th in SE, or 3rd, 5th and 7th in PP. Pentodes preferred for balancing purposes; triodes will drive each other and the total effect is harder to evaluate (not necessarily dangerous, just different).

Tim
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.