Resistors on OT primaries

[ron vogel]

Hypothetical question:

Say I have a 6.6K primary impedance on a PP output transformer (15ish watts). I want closer to 8K.

Can I just add a 500r to 1K resistor on each primary, and what kind of watt rating should I be looking at?

[eschertron]

Hypothetical question:

Say I have a 6.6K primary impedance on a PP output transformer (15ish watts). I want closer to 8K.

Can I just add a 500r to 1K resistor on each primary, and what kind of watt rating should I be looking at?

What is your intended result? A quieter amp? An amp/space heater combo? The resistors probably will do nothing for you except to suck up some available power. If this is a hypothetical test rig, I'd say go for and see how the resistive loading affects the sound. You'd probably need a resistor on each leg rated for at least the full dissipation of the tube to which it's attached, so 750R at 10W for each?
I've seen schems with resistors on the plate instead of the screens (Vox AC15 1959 drawing), and have always assumed it was a misprint (accidentally or deliberately). Maybe someone else can chime in with the answers.
Now that I'm thinking about it, I've wanted to build an amp with the screens at a higher voltage than the plates; this may not be the way to do that, though.

edit: If, on the other hand, you actually want to build a gigging amp, I'd say redo the load lines for the OT that you have, and adjust other parameters to fit, rather than shoehorn extra components into the circuit.

[Chuck H]

Resistors in series on the OT primaries will add considerable power loss and sag. As mentioned, it would be better to adjust other parameters. Probably the easiest way to bump the OT's reflected impedance on the power tubes would be to change the speaker load. Sort of tricky changing from 6.6k to 8k though. What is the secondary load rating and what is the nature of the speakers for this amp (ohmage, wiring, is this a head or combo, etc.)?

[R.G.]

Say I have a 6.6K primary impedance on a PP output transformer (15ish watts). I want closer to 8K. Can I just add a 500r to 1K resistor on each primary, and what kind of watt rating should I be looking at?

Eshertron is correct - sure you can do this, but why would you? There are several possible reasons.

One is for a test rig, to see what the voltages and currents are. That's OK, but they will not be the same as in a real 8K P-P transformer setup. Pure curiousity is another reason, and a really, really good one. Tinkering is always a valid reason to try something. It's how we find things out.

If the idea is to use a 6.6K P-P transformer with tubes that are rated for 8K P-P in hopes of getting the proper matched load, there are some issues. From a technical standpoint, the only reasons to use a given impedance ratio from primary to secondary are to (1) get the most power from a given power supply and set of tubes and (2) get the widest frequency response possible from that transformer, both of these at the same time.

The reality is that the resistors in series wth the primary (or in parallel, amounts to the same thing) eat some of the available power from the tubes that might have gone into the transformer and hence into the load. So the tubes may be working at a better power output for the given power supply, the extra power produced by the tubes is partially eaten by the added resistors and doesn't get into the transformer, and then the load. So it's useful for testing tubes, but you might do as well with just resistors, no transformer at all.

You *can* correct the apparent loading on the plates by increasing the secondary load. You can put resistors in series with the secondary load, and this is transformed back into the primary. Transformers don't have impedances, they have RATIOS. This has the same problem as resistors in the primary, though. The resistors you add get the extra power, not the load (presumably speakers).

There was a lot of work done on matched loads for tubes and speakers back in the 40s and 50s. This led to some useful charts. If you plot output distortion on a secondary versus plate-to-plate load for a given set of tubes, there is a broad, flat hump of lowest distortion at some impedance between the plates. If you plot power output versus plate-to-plate load, there is a broad, flat hum at some impedance. The humps peak at close, but not the same, impedances. For a 6L6, the lowest distortion is about 6.6K, but the highest power is about 4.4K. Hifi guys run their 6L6s near 6.6K, guitarists run them at 4K to 4.4K. For EL84s, the lowest distortion is about 8K-ish, the most power a bit lower.

The important thing to remember is that mismatching the tube best-power or best-distortion points doesn't change the output dramatically. It lowers the power or raises the distortion a bit. Given the human ear's high compression of loudnesses, the loss/gain in apparent volume doesn't change a whole lot for 2:1 changes in power (for instance).

So changing the P-P impedance from 6.6K to 8K isn't all that useful for an actual working amp. It may well be for gathering data or satisfying curiousity.

[ron vogel]

Basically, it's for a tonal change. It would probably easier to just use the 4 ohm taps to double the impedance, but I want to fine tune it. I have an amp that started as a 5F11 and now it's pretty far removed from that circuit. It "sounds" better biased hot, but the circuit works better biased cold. I need the output tubes to work harder to get the sound I want, but also bias colder.

The right way to fix where I'm at is to replace both the PT and OT, but it's sooo close where it's at that I'd like to push my progress a little further.

The amp is used soley for recording, and because of the ambient environment and recording gear I use to get a certain sound...I need it to be set-up a certain way.

If I have to get a new PT and OT, then it's going to open up a whole new can of worms. In the back of my mind I have been thinking of going to a long tail PI, and dropping in a BF bassman tone stack too...so I'd be adding another tube. As much fun as that would be, I'm putting off some recording projects because of me tweaking this thing.

[soundmasterg]

A question R. G.....why would a designer want to use series resistors in a given design? For instance, the Sunn 2000S and the Vox AC100 use series resistors with the Sunn using 47 ohm on plate and screen in an UL setup, and the Vox using 100 ohm on the plates. It seems to me that restricts power, and it makes me realize why an AC100 project build I was working on here a couple years back was only making 70 watts instead of the more typical 100 it should have been....at least I think that may be a contributing factor. Any comments?

Greg

[ron vogel]

Yeah, from the replies in this thread, I can see where this is going. The power loss I can deal with; but I have enough sag as it is.

It's a PP 6V6 amp, but 6K6's sound better in it. Problem is the 6K6's bias so hot at the voltage they are seeing (355-360) that it kills the trem circuit. I have the trem working great with the 6V6's biased cold, but then the preamp tubes voltage is too high for a nice tone and the amp gets boring.

Basic rundown. 6.6K OT with 4-8-16 ohm taps. Running a 8 ohm speaker. With 6V6's I see 385-390 biased cold, and 350-360V biased hot.

I see 200V at the preamp biased hot, and 270v biased cold. I have made so many changes to this thing it is a mix of 5F11, 6G2, AA1164, and 6G3...lol

[eschertron]

I'm not sure what you mean "the circuit works better biased cold". The colder you bias it, the less compression you get at peak voltage (lower off the curve of the knee). This is the same operating point you'll get with an OT that has a higher primary impedance. If you like the sound of the amp biased hotter, I'd predict that you want a lower primary impedance at a given bias level, rather than higher.
Of course, this is a good hypothesis to test with a rig that can adjust the reflected primary impedance to any value! I like R.G.'s suggestion to put the resistors on the secondary side. It's done all the time by people looking for loudness control. Not perfectly transparent, but maybe close enough to tweak for your purposes.

edit: I see your reply about the trem.... I get it. Thanks!

[Chuck H]

You could always just bias a little cold and increase the resistor value in the HV rail that feeds the preamp. Easy peazy.

If power loss isn't a problem you could use a resistor across the speaker load. This will reduce the dynamic play between the speaker/s and amp a little but I doubt you would notice. So, if you used an 8 ohm speaker on the 4 ohm tap and then parallel a big 20 ohm resistor with the load you'll get pretty close to the higher reflected primary impedance on the tubes that you want. Likewise you could use a 16 ohm speaker on the 8 ohm tap and parallel a 33 ohm resistor. If the speaker load is up for grabs you could skip the resistor and get creative by using a 16 ohm and an 8 ohm speaker or even a three speaker configuration. A 12 and a ten or a 12 and two 10's perhaps. That would also add mic'ing options in the studio for recording to get slightly different tones. It won't be exactly the 8k reflected impedance you're asking for but it'll be closer and won't induce the sag that resistors on the primary side would.

Important to note... I don't know if that amp employs NFB but if it does the change in impedance will alter the loop gain and change the tone of the amp. So the loop would need to be adjusted if you want to keep things with the rest of the amp created equal.

[potatofarmer]

A question R. G.....why would a designer want to use series resistors in a given design? For instance, the Sunn 2000S and the Vox AC100 use series resistors with the Sunn using 47 ohm on plate and screen in an UL setup, and the Vox using 100 ohm on the plates. It seems to me that restricts power, and it makes me realize why an AC100 project build I was working on here a couple years back was only making 70 watts instead of the more typical 100 it should have been....at least I think that may be a contributing factor. Any comments?

Greg

I'm not RG and I can't speak for the Vox but I've been reading a lot about UL lately. In the Sunn example they added small "plate stoppers" to suppress oscillation. Same rationale as (small) grid or screen stoppers. There's a thread about plate stoppers here: http://music-electronics-forum.com/t22558/

[ron vogel]

The trem was tricky to get right, and after all said and done made the board look pretty ugly. What you are saying makes a lot of sense compared to what I am seeing though. Transients really suffer when I bias it hot. Clean appregios sound great, but that first power chord strike lowers the volume dramatically.

I was looking in the direction of a higher impedance only because of the voltages I am dealing with are more in line with the silverface era, and the OT on the later Princeton's are 8K.

How hard will it be on the tubes if I run 3.3K ? I may try it and bias it on the cold side. I basically want the output tubes to have more current/voltage draw than the rest of the circuit to balance it better. If there's another way to approach it, I'm all ears! Seems every other avenue I've gone down with this has had too many drawbacks.

[ron vogel]

I may just try that on the rail. Looking back, the only reason I didn't was because of the PI. I tried many things on the PI (including biasing it with an LED), but the Stokes mod worked the best. If I could get the preamp voltage to stay between 200-220V with a colder bias on the 6V6, it might get me in the ballpark quick. The second gain stage cathode bypass and PI relationship was a treat to get right, and I understand why Leo didn't use one in this circut. From what I have learned with this thing so far, I'm a little less scared of messing with the rail voltage now.

I have an adjustable NFB. Goes from 2.2K to 500K. The trem worked best for me w/o NFB, so I have a switch on the NFB pot. Lower than 2K the NFB was too much, so I started at 2.2K, and turns up to 500K until the pot switch engages the trem.

[ron vogel]

On a side note, this amp sounds terrific and quiet. I'm just adjusting it to suit my recording needs and playing style. I'll have to get a schematic up of it when I'm there...and I almost am.

Here's some of my crap music to get an idea:
http://soundcloud.com/ron-vogel/coming-home
http://soundcloud.com/ron-vogel/musical-chairs

[ron vogel]

If you look at the b+ resistor (10k) on the 6g2 circuit...I raised it to 15k according to the above recommended change. Worked well. I ran out of time to mess with it tonight, but I'm going to try 22k. Seems to do the trick...btw, this amp is freakishly loud!

[Chuck H]

If you look at the b+ resistor (10k) on the 6g2 circuit...I raised it to 15k according to the above recommended change. Worked well. I ran out of time to mess with it tonight, but I'm going to try 22k. Seems to do the trick...btw, this amp is freakishly loud!

If it's making the recordings sound like you want them to, it's exactly as loud as it should be

On a side note, I feel your pain WRT dialing in bias wiggle trems. I had to do it recently and found that the amp I employed it on isn't really suited to this style of trem, BUT, it's really good when it works. I was suffering from that garbled crossover hash at the bottom of the wiggle on all but the cleanest tones. For another reason I had to swap out five different pairs of power tubes in this amp. I tried each pair at different bias levels to find the best compromise of all the amps qualities and what I discovered after some observation was that a little mismatch seems to clean up that garbled distortion. Not so much that the amp hums or fails to produce it's intended power. Just a little and all is well. And, same as you, I found that less NFB helps too. Not an option on this amp unfortunately. I'm putting a note inside the chassis as a reminder to myself for whenever it comes in for service Who knows, my memory may be failing by then?