Output Load Line

[Tage]

I am building a bass amp for a friend, he want's something with 4 KT88's going for between 150-200 Watts. He want's it to be basically identical to the orange OD120, with a few tweaks to suit a bass. I have decided to run with a B+ of 520V and an output Transformer primary impedance of 1.5kOhms.

My issue is drawing a load line to try and find a good bias point for the output stage, I have been reading (and have ordered the book from) the Valve Wizard page but when I try to draw a load line like his one in his Push Pull output stage mine is way off the chart. Any one know were I might be going wrong here?

[MWJB]

Design the bias supply & trim pot to sweep to around -65vdc/-70vdc at its most negative, bias to 55mA per tube max, by plate current.

[Steve Conner]

Yes. By choosing the B+ and primary impedance you have already set the load line, so just bias it to whatever point gives a comfortable idle dissipation.

[Tage]

Sounds good, 1.5k is a pretty decent lad for these tubes is it not? I just had it reccomended by Paul from Mercury, big budget on this amp

I would like to figure out how to draw a class A/B load line, but maybe I will do that later on, I need to get my stuff together now. Also does anyone think it would be better to go with a higher B+ of say 560V? I know of some other bass amps that run this such as the traynor YBA3

[Steve Conner]

In a high-powered Class-AB amp, the value of B+ that matters is what it sags to under full load. In that context, the difference between voltages of 520 and 560 measured at idle isn't worth worrying about.

I would rather have a power supply that was 520 at idle and 470 at full blast, than 560 at idle and 400 at full blast.

If you want 200W from four tubes, then I recommend the old GEC datasheet conditions that produced 100W per tube pair. (See page 2 here: http://www.drtube.com/datasheets/kt88-mov74.pdf) Aim for something like 560 fully loaded, and the idle voltage will be over 600. Use 4.5k a-a load per tube pair (2.25k for 4). Note the screen voltage of only 300V.

Cross your fingers and hope the new ones are as good as the old GECs. If they're not, then the 600V supply and 2.25k load will give you a better chance of meeting the 200W goal, than 1.5k with a lower B+ would.

Maybe Jon Wilder ("Wilder Amplification on the forum") will swing by and tell you how to draw a Class-AB load line Or you could try searching his old posts.

[TransLucid]

To draw the load line for a PP stage, you need to use two characteristic curves, butted along the Ip=0 line, one upside down, joined at the (same) power supply voltage. The load line goes through that point and has a slope equal to the OT's primary impedance. It's a little hard to imagine from this description, but you can find these charts in many references, including my book. It's useful for ensuring that the tubes stay in the safe operating area, among other things.

[Tage]

So I feel alright with how I will bias the amp, but now I have a question about using a higher B+ in a circuit designed for about 490V (optimal, not under load) if I increase that to say 560 do I need to increase the size of the resisters that drop the voltage for my pre-amp stages? the schematic I am working from shows them as a 22k followed by a 10k, although other schematics show two 33k resistors.

[ThermionicScott]

One way to determine the ideal load without characteristic curves is to play with different values for your dummy load. The load that allows your amp to make the most power will be the one that is (closest to) intersecting the kneeof the curves. From that, you can determine the impedance ratio so the end user doesn't need to have a goofy speaker load.

-Scott

[Tage]

I'm going with the Partridge 200watt transformer through Mercury. I noticed on there schematic that they have 22k grid stopper resistors, while Sunn amps like the model T only use a 1k resistor. Is this a crucial value will 22k suit bass guitar better than a 1k or vice versa?

I hope these questions aren't too vague, I'm designing something that is starting to seem like quite a jumble, but I hope it isn't really.

[TransLucid]

The screen grid has a maximum specified voltage, which is in some cases lower than the maximum plate voltage (eg, for EL34s). If you intend to run the plates at 560V, check to make sure that the screen grid can handle that voltage.

Second, larger screen grid resistors (in addition to dropping more voltage) protect the tubes from excessive screen grid dissipation. You can certainly go higher than 1k. Using 22k may result in high power dissipation in the resistor - make sure you calculate wattage in any event. Don't forget, though, that changing the screen grid voltage changes the tube characteristics.

As far as the preamp tubes, they also have a maximum plate-cathode voltage spec, so you shouldn't exceed that rating. Normally, you drop more voltage in the power supply by increasing the values of the series resistors (again, watch the power dissipation).

[Tage]

Sorry the 1k and 22k grid stoppers I was refering to were for the control grid not the screen grid, I think I will use 2k screen grid resistors to try and get my screens at around 300V

[GregS]

Would using excessive screen grid resistance introduce a sag or compression effect that might not be desirable in a bass amp?

Greg

[ThermionicScott]

Using resistors to drop from 520V to 300V does sound excessive (lots of heat and wasted current). Are you just trying to get things to the point where you can draw load lines on published curves?

One technique that might get you in the ballpark would be to "pretend" that your plate voltage is also 300V, as if you were using a low-voltage supply and a choke between your plate and screen supplies. Draw the loadline that intersects the knee nicely, and then use that load with your 520V supply, not doing anything fancy with the screen supply apart from suppressing oscillation. This method becomes less accurate the further you are from the published curves, but it's a start. (Since Genalex/MOV was kind enough to publish curves for 100V, 150V, 200V, and 300V operation, you can try this technique on all of them and note that the ideal load keeps dropping as you increase the screen voltage. At 300V, I came up with something like a 2.4K plate-to-plate load for two tubes -- it would probably be even less for two tubes at 500V.)

HTH,
- Scott

[pdf64]

Regards grid stoppers, note the 100k grid resistor limit for high dissipation arrangements, on the linked data sheet post #5. So with a 22k grid stopper, your grid leak will need to be less than 78k.
But 22k seems an unnecessarily high value - maybe they were having to mitigate for poor lead dress or excessive global feedback. Pete.

[J M Fahey]

Would using excessive screen grid resistance introduce a sag or compression effect that might not be desirable in a bass amp?

Yes, definitely.
You can make a 600V supply with a center tapped transformer , a bridge, and two capacitors in series.
The transformer CT which is connected to the union of those caps is an excellent 300V source.
Since you'll use capacitors in series anyways, it's the best optiom.
Or you can make a regulated and adjustable 300V at relatively low screen currents with a Mosfet.
Optimum load calculation:
From your .PDF, see graph 1473, at the bottom of page 7.
The best load line is the one which opotimizes voltage and current swing, meaning highest power.
Load line for *one* tube, one half cycle (the other half is symmetrical, so we can get away with this simplified calculation).
Transformers do not transmit DC voltages or currents, so we must calculate their *variation* (called Delta) which is what we will be able to transmit to the speaker as audio.
The chart stops at 500V; we may use a 600V power supply (at idle) but it will certainly drop.
How much? Won't really know until we already have it on the bench, but I can guess that +520V might be a reasonable value for a *good* one.
Minimum voltage reached by the plate? The curve bends, has a knee, but I see 80V reasonable, being that it still can supply hefty 400mA at that voltage, at Vg1=0.
So useful plate voltage variation=Delta V=520-80V=440V.
Now Delta Current:
Current at idle= 55mA (you suggested it, sounds reasonable)
Peak current=400mA
Delta current=400-55mA=345mA=0.345A.
Plate load impedance for *one* tube amplifying half wave= "Delta V"/"Delta C"=440V/0.345A=1275 ohms.
For both tubes in push-pull, Plate to Plate impedance is 4x that, so you need a transformer with a primary impedance of 4 x 1275=5100 ohms. for *two* KT88 in push-pull.
For *four* tubes in push-pull impedance halves= 5100/2=2550 ohms.[/B]
This is following published data for original KT88; you start your design with them, and then make empirical adjustments to optimize power or sound.
Those 2250 ohms suggested above are quite close to this; I am sure that Wilder's graphical calculations will be close too, if based on the same set of curves; Mercury's suggestion of 1500 ohms seems too low and unrealistic.
The fact that "Partridge used that" sounds more like hearsay or Mojo than actual calculation.
Of course I may be wrong.

when I try to draw a load line like his one in his Push Pull output stage mine is way off the chart. Any one know were I might be going wrong here?

I guess now you know why.