Orange OR-120 clone. Adjusting Split load driver for 2 x EL-84?

[Gingertube]

To reinforce what Juan says, If you want to be able to overdrive the output stage then the cathodyne splitter is NOT the correct choice. Use the differential splitter (Schmitt or longtailed pair) or even an old fashioned paraphase instead.

I'm not going to rehash the debate about the output impedances of the cathodyne at the anode and cathode when the anode and cathode loads are equal. Its been done to death, particularly over at DIYAudio. The take away note from all of the theory and modelling which has been done, is that those anode and cathode output impedances ARE equal (approximately = 1/gm) while the loads are equal (due to feedback effects) but they rapidly become VERY unequal as soon as you drive hard enough for output tubes to start to conduct grid current.

The advantage of the Cathodyne splitter for HiFi (when you are not driving output tubes into grid current) is the exceptionally low output impedances which allow driving of the output tube Miller capaitance to higher frequencies. This is something which is not required in a Guitar Amp.

Using an inappropriate circuit and then trying to fix its "warts" is just bad design.

It is significant to note that the next model of that Goldentone I referenced, abandonned the Cathodyne Splitter and used a Long Tailed pair instead.

Cheers,
Ian

[frus]

Well just to stir the water some more 😁 I've built a 5f4 tweed super (cathodyne pi, 6L6) a long time ago and tried various things to 'improve the bad design'. They mainly worked as advertised, but one thing is really mind boggling. The stock circuit (btw, it does use 56k resistors on cathodyne with 6L6's which are hardest to drive) sounds totally great and 'normal' with old-production power tubes (I tried jan/philips, original tung-sol 5881, rca, ge, sylvania), they all sounded pretty similar. However, new production tubes (jj, new russian tung sol, sovtek) sounded splattery, blocking distortion-y, in a word what we would expect from the circuit. In scoping I only found that the old tubes go into cutoff more gradually, but nothing more. Anyone have an explanation?

[pdf64]

Perhaps the g1-k diode characteristics (especially as it transitions into forward bias) of newer production are different to vintage production?
They don’t seem to be a specified characteristic, I wonder if they’re noted anywhere, eg in the type approval documents?

[Helmholtz]

The take away note from all of the theory and modelling which has been done, is that those anode and cathode output impedances ARE equal (approximately = 1/gm) while the loads are equal (due to feedback effects) but they rapidly become VERY unequal as soon as you drive hard enough for output tubes to start to conduct grid current.

Sorry, this is wrong as I wrote before.

Yes, there is feedback, but it is not the typical voltage feedback which would decrease anode output impedance. Rather the feedback voltage is developed across the cathode resistor by action of the load current. This is current feedback which in fact increases anode output impedance.
The anode impedance is 50 to over 100 times higher than the cathode impedance depending on anode resistor value.
Consequently the cathode is able to drive more load current than the anode before its voltage distorts.

Here are the relevant formulae:

triode.pdf


The actual output impedances can be measured using an LCR that allows R measurement at 1kHz, using a DC decoupling network consisting of a 100n capacitor in series with a 1M resistor from each output to ground and connecting the meter across the resistor after the cap is fully charged. I often do such measurements of internal impedances.

Alternatively you could wire additional resistors from the outputs to ground and calculate internal impedances from the signal voltage drops. According to theory, output voltage will be half, if the external resistor equals the output impedance.

[Malcolm Irving]

Does Merlin's chapter help to clarify? (See Section 12.1.2):

http://www.valvewizard.co.uk/cathodyne.pdf

[Helmholtz]

Does Merlin's chapter help to clarify? (See Section 12.1.2):

http://www.valvewizard.co.uk/cathodyne.pdf

Thanks, this completely confirms my above statements.

There are more books discussing the cathodyne aka Concertina. Highly recommended: Richard Kuehnel's book "Guitar Amplifier Power Amps".

[fiveightandten]

Thanks for the detailed information everyone, and apologies for the late response. Unfortunately DIY project planning has been taking a back seat to life obligations.

Pity you chose an Orange and not any other Guitar amp out there.
Main problem when building "EL84 versions of large amplifiers" is that EL84 are way easy to drive (so easier to over-drive) than 6L6 or EL34, so amp response changes ... a lot.
That can reasonably be compensated for in LTPI such as in Fender/Marshall (and most anybody else) by using slit plate load resistors, so if , say, you want to halve available drive signal you replace 100k plate resistor with 2 x 47k in series, the 82k one with 2 x 39 or 43k ones, and in both cases take signal out from the centerpoint of each series, effectively halving drive available to power tubes.

But Orange uses a cheesy Cathodine PI (you call it split load but thatīs the wrong name) which does not allow that.

Just as an experiment, I would suggest you attenuate signal reaching power tubes, build circuit as-is, but use properly biased and loaded EL84 as outputs and attenuate signal reaching grids 3:1 , which is about what you need.
You can add 470k resistors in series with power tubes coupling caps C24 and C23.
Then on top tube(s) you will have C24+470k > R22 to ground, and on the bottom tube(s) C23+470k > R23 to ground , both nets providing 3:1 signal attenuation but leaving biasing network as-is.
I am referring to the first schematic, just to avoid confusions.

They used a few different versions, so whatever might be the best starting point to alter:


Not the same, *always* make clear what are you talking about.

Not too important, if at all, and you are starting with lower +V anyway, so donīt obsess with that , in any case sound will NOT audibly change unless you get to starvation voltages, say 90V or less, not the case here.

Thanks for the info! So, I suppose we'll call this approach #1 (series resistors to attenuate the PI output signal). This makes sense and if it works, would be a very simple band aid on the issue.

I was planning on fixed biased EL-84's though. Are there issues with feeding the bias voltage into that R22/R23 junction with the added series resistors present?

I understand the adverse response to the Orange circuit design, as it is unconventional. Though, I do love the raw jagged sound of these old amps overdriving. I own/built low powered Vox, Hiwatt, and Marshall type builds, and 15W 2 x EL-84 is the perfect amount of power for me these days, but no-one makes this old Orange circuit coupled to such a power section, hence my want to build it.

The cathodyne is perfectly fine and used with great success in a whole bunch of amps. I agree that the DC coupled setup in orange and matamp designs is going to flap around like crazy but that is where the tone comes from in these amps! It may not be your cup of tea but if you want that sound that's the way to get it. AC coupled is perfectly fine and in some high gain circuits you don't want the LTPI flavour in the tone.

I've played around with orange-esque circuits using a pair of 6V6s at ~300VDC which may be better suited to this circuit than EL84s as they don't immediately bounce off the rails when presented with a tickle on the grids. The first thing to try is to lower the cathodyne plate and cathode resistors to tame the signal swing. 47k is more reasonable. Are you dead set on EL84s?

I'm not dead set on it, though my thinking was that the amp gets its sound from the output tubes working, and likely won't sound right without that. I have a few of these, one OR-80 "Overdrive" model with the PPI master in it, and it just plain sounds terrible unless the master is up full. I know that a pair of 6V6's working is typically too loud for my situation.

My plan was to buy a used OR-15, gut the chassis, and wire the OR-80/120 circuit up using the O-15 transformers. I could easily add a choke if needed, and would need very little chassis modification to mount things. A 6V6 build would change that plan for sure, and I fear I wouldn't be able to get the amp working, even at 300VDC. But I defer to the knowledge here, as I have precious little experience with theory and design, and up until now I've just built clones and done repair work.

I'll also add that I've built a few orange circuits and I've played a bunch of originals and a couple of Matamps GTs too; I like how they sound when other people use them but I just can't get them to work for me. I generally like that wet sock (tm) stoner doom tone too haha!

The trick is keeping the bass in check. They fuzz out a lot with high FAC settings. It's a sound, for sure. But I prefer a classic crunch out to them, with the characteristic overdrive they have.

If the main purpose of the project is to build a lower wattage OR, why change output tubes at all? Lower wattage comes from lower voltages not from smaller tubes. EL 34s work fine at 250-300V. Transconductance will be lower at lower screen voltage but probably still be higher than 6V6s'. The lower gain can probably be partly compensated by a higher load impedance. Lower voltages give more freedom to choose a load line.

Will EL-34's provide the same kind of saturation and dynamic response characteristics running at such low voltages?

For what it's worth, I've built an OR clone that runs a pair of 6V6s, not that different from EL84s, and didn't make any changes to the circuit to reduce the signal that drives the power tubes. All it does is shift the point on the volume knob where distortion takes hold, and affects the amount of power tube distortion that you can achieve.

I did however follow Merlin's recommendations for reducing the possibility of nasty distortion from the cathodyne and added a large grid stopper resistor to it.

It's also important to note that there are two main versions of the OR-120, with the main difference being the cathodyne. I feel that the AC coupled cathodyne is superior to the DC coupled version.

Thanks, do you mind sharing what your plate voltages and PI setup were in that amp?

The cathodyne PI - a copy of something I posted elsewhere regarding an Aussie Goldentone (6DQ6 Output tubes).

The reason for the 100K anode and cathode load resistors on the cathodyne for big amps is purely for ease of swinging the required output tube grid drive voltage. The required voltage swing for the Goldentone is about the same as what is required in the ORANGE MATAMP to drive a quad of EL34. It uses 100K too. All of the EL34 and 6L6 Amps I've seen with cathodyne splitters use 100K or 82K.

Smaller Amps with EL84 or 6V6 which do not require the same voltage swing, typically use 56K. That is, they are swinging the same current in the cathodyne as the bigger amps but generating less voltage drive into the output tubes.

Calculating Overdrive Effects
If MATHS gets on your wick then skip to the SUMMARY below:

For equal loads on Anode and cathode on the Cathodyne

Zout = RL.ra/RL(u+2)+ra

The ra term on the bottom line is insignificant compared to RL(u+2) term so drop it. Then the RL terms top and bottom lines cancel leaving

Zout approx = ra/u+2

At typical values of u (>=20) u+2 approx = u, so simplify again

Zout approx = ra/u = 1/gm

So for equal loads at the anode and cathode the Zout at both is approximately 1/gm or about 650 Ohms for a 12AX7

If driving output stage directly The equal loads on Anode and Cathode will NOT be guaranteed if:
1) Output stage strays out of Class A (When a tube cuts off it has no gain so Miller capictance will change, particularly with Triode Mode Output, less so with Ultralinear and less so again in Pentode Mode)
BUT more importantly
2) Output Stage strays into grid current (trying to overdrive the output tube). When this happend the load resistance presented by the output tube grid drops significantly.


If the Anode load drops significantly then:

Zout cathode = RL+ra/(u+2) x ra/RL The ra/RL term insignigicant so

Zout cathode approx = RL+ra/u+2 At usual values of u

Zout cathode approx = RL/u + ra/u = RL/u + 1/gm

That is it increases by RL/u


If the cathode load drops significantly then:

Zout anode = RLxRL(u+1)+RL.ra / RL(u+2)+ra

RL squared (u+1) is much larger than RL.ra and RL(u+2) is much larger than ra so

Zout anode approx = RLxRL(u+1)/RL(u+2)

and at reasonable values of u

Zout anode approx = RL


SKIP TO HERE:
SUMMARY:
As the loads on Anode and cathode become unbalanced (as when overdriving the output tubes) then

Zout anode increases from 1/gm toward RL (from 650 Ohms toward 100 KOhms in the Goldentone)
Zout cathode increase from 1/gm by maximum factor of RL/u (from 650 Ohms toward 650 + 100K/100 = 1650 Ohms in the Goldentone)

The imbalance from an over driven output will therefore be worse with the 100K but the difference between the anode and cathode output impedance in overdrive is so dramatic it doesn't really matter if its between 650 Ohms and 100K or between 650 ohms and 56K, Also the effect of this will be asymetrical, affecting one side of the push pull and therefore introducing large gobs of 2nd harmonic distortion.

The trick here is to use large grid stops on the output tubes to reduce the loading effects on the cathodyne. Take those output tube grid stop resistors way up to say 47K. THis is effectively tuning the amount of 2nd harmonic distortion you get in overdrive.

Is this post useful or should I pick up my marbles (if I can find them) and go home.

Cheers,
Ian

Thanks Ian. It's extremely helpful, thank you. I can't say that I follow all the math there, but the theory behind is does make sense. So, to clarify, are you suggesting to lower the PI anode and cathode load resistors to 56K *and* raising the grid stoppers from 2k2 to 47k? What are the benefits of this approach compared to series resistors on the PI output (470K as Juan suggested), or splitting the anode and cathode load resistors and taking the output from the middle (as Malcolm suggested)?

To reinforce what Juan says, If you want to be able to overdrive the output stage then the cathodyne splitter is NOT the correct choice. Use the differential splitter (Schmitt or longtailed pair) or even an old fashioned paraphase instead.

I'm not going to rehash the debate about the output impedances of the cathodyne at the anode and cathode when the anode and cathode loads are equal. Its been done to death, particularly over at DIYAudio. The take away note from all of the theory and modelling which has been done, is that those anode and cathode output impedances ARE equal (approximately = 1/gm) while the loads are equal (due to feedback effects) but they rapidly become VERY unequal as soon as you drive hard enough for output tubes to start to conduct grid current.

The advantage of the Cathodyne splitter for HiFi (when you are not driving output tubes into grid current) is the exceptionally low output impedances which allow driving of the output tube Miller capaitance to higher frequencies. This is something which is not required in a Guitar Amp.

Using an inappropriate circuit and then trying to fix its "warts" is just bad design.

It is significant to note that the next model of that Goldentone I referenced, abandonned the Cathodyne Splitter and used a Long Tailed pair instead.

Cheers,
Ian

It seems that thatch 22 here is overdriving the output stage is the characteristic sound of the original circuit. It seems that there are a few approaches that may work to try and couple a lower powered output section to this preamp and PI, but it may not yield something that sounds good, let alone like a quieter version of the original amp.

[dwrockdoctor]

This is an epic necropost but thought I'd share that I did this exact thing but to a Vox NT15H Night Train head that has EL84's. Because it already had a cathodyne phase inverter I basically went to work replacing and cutting out any components that didn't line up with the OR120 preamp and phase inverter circuitry. I basically left the power supply to all the preamp and phase inverter tubes intact and I didn't change any components after the phase inverter, just anything upstream in the preamp that didn't correlate with OR values got changed and I cut out and rebuilt the whole EQ section (basically wired the James/Baxandall circuit to the actual tone pots themselves).

It's turned out great, gets that old orange/matamp grind at a more sane volume. I actually put 6V6's into my GT120 and adjusted the bias to see how that fared and though it sounds great it's still damn loud and I had the Night Train given to me and found the amp is overly bright and fizzy so thought it would be a fun experiment to convert to old OR specs. It has some really solid iron for such a small amp. My dB reader showed an average of 105dB with the 6V6 matamp vs 100dB with the EL84 modded Night Train/OR which I think is a pretty decent reduction in overall volume while still getting that end of the world OR/GT tone. It's a bit bright, maybe due to the EL84's and how the power amp circuitry is voiced but that can rectified just fine by turning down the treble a decent amount. I might look into some mods for the power amp section to lessen the brightness at some point. Otherwise, it worked out fine. Hope this info helps!