Loadline Tools

[Cressy Snr]

There is a calculator on this site that calculates distortion https://www.vtadiy.com/loadline-calcula ... alculator/
This can be used in conjunction with the calculator Paul has illustrated above. You can find a suitable op point and self bias resistor with the first calculator, then open another tab, transfer the figures into the other calculator. You input the max swing available and the software will fill in the distortion figures for 2nd, 3rd, and 4th harmonic below.
My pic is a bit big, as I have no way to resize it and iOS saves screenshots as PNG, which can’t be resized either for direct upload to the board. So you’ll have to put up with my giant picture for now unless Nick can do something about it.



So...yesterday I found that although I could get 5W with a 2K5 load, the distortion performance was risible. 8% THD at full power and it never going below 3% no matter how narrow the swing.
The KT77 therefore does not produce hi-fi with a 2K5 load no matter how much one tries to spin it.

A 5K load OTOH does get good performance, as can be seen above but the power is obviously halved. I’m lucky I’ve got the older type Hammond 1627SE with multiple secondaries, so coupling is not compromised with an 8Ohm speaker on the 4Ohm setting.

To get back to a higher power output, with decent distortion performance would need a KT88 or above... or....a 300B.

Maybe I’ll have to bite the bullet and finally build an amp with 300Bs. You lot know it makes sense but sadly it is not an affordable option, so I’ll continue the beer budget experiments. KT120 looks a good, future bet but certainly not with a 6SL7 in the driving seat.

[JamesD]

Umm - anything but a 300B!!!

Even the JJ 2A3-40!



I assume you are going for KR300BXLS - you know a proper 300B

Alright if you want beer budget but similar to 300B output and curves.... then GU-50 in triode is very well proven - lots on the web about them

[Paul Barker]

I’m pretty sure Steve has talked about GU50 quite some years back when he was into Pentode designs with local feedback; looks on eBay like they’re still cheap and readily available.

Steve the relationship between power and load when studying operating points isn’t linear or simple so as to say power doubles or half’s. That is because as you alter things you are planning to shift the loadline to a place where you can continue to drive it with greater amplitude yet remain within listenable music reproduction. It’s about operating the valve optimally so you get most listenable power.

[Cressy Snr]

Umm - anything but a 300B!!!

Even the JJ 2A3-40!



I assume you are going for KR300BXLS - you know a proper 300B

Alright if you want beer budget but similar to 300B output and curves.... then GU-50 in triode is very well proven - lots on the web about them

Cheers James. You’re not a fan of modern production 300Bs then.

Steve the relationship between power and load when studying operating points isn’t linear or simple so as to say power doubles or half’s. That is because as you alter things you are planning to shift the loadline to a place where you can continue to drive it with greater amplitude yet remain within listenable music reproduction. It’s about operating the valve optimally so you get most listenable power.

I managed to set up the calculator, so that with 5K, I got just under half power for the best distortion I could manage. Of course that is not to say that a higher quantity of 2nd wouldn’t be listenable. I think (I’m often wrong) that with the use of this tool I can correlate what I hear with what the models are saying.

How ‘loud’ things sound seems to be directly proportional to the amount of 2nd harmonic present in the output. And if you can arrange for a decaying spectrum 2nd to 3rd, to 4th and so on, it all sounds much better to the ear. But then we know that already. It is however the first time I, personally have been able to see/ hear it in action.

With 5K load, and less distortion this ‘loud‘ ‘exciting’ behaviour is much reduced, but then we have to watch out for clipping, due to the reduced power which is not good either. A juggling act where you optimise one parameter at the expense of another.
2.25W is fine into 100dB/W speakers in a small room. That’s my story anyway.

It would seem from the availability of these models, that ‘build and hope’ is a crap strategy all round. Oh that these models had been available 15 years ago, I wouldn’t have been forever frustrated. I’m not one for giving up, but bloody hell 30 odd amps is two per year for 15 years! It provides good copy but it’s not good for the nerves of the one providing the copy. I for one am grateful that people provide these tools. At least now, I have the facility to check things out and order the right components, before getting out the soldering iron.

[Paul Barker]

Steve, you’re image views on this forum on an iPhone ideal presentation.

I’m noticing though these days less and less software designers are bothering with iPhone functionality.

What mostly bugs me is the portal I have to fill in for my work I get from Corgi (most of my work). Works a treat on a Microsoft pooter and only works on an iPhone with a Sh1t load of “workarounds”! A Workaround is the name for a load of complexed ways of arriving at functionality which makes pen and paper look like it’s on the way to a come back.

[JamesD]

Its not so much that I'm not a fan as:
1) 300B are not easy to use for low distortion if you want more than about 2W output from them and if you don't why aren't you using a 2A3 or PX4 or something similar
2) They are not good value just because they are 300B they cost more...
3) They start grid current and lowering input impedance early - well before Vg=0V so hard to drive
4) They actually have a low heater power for there output - this affects 1 & 3 -less heater power than a 2A3 for instance.

Having said that their curves are very nice so if you design around the points above it can work nicely - also not all 300Bs suffer from the above as many different designs and implementation are called 300Bs as the curves are close so its pick and choose what you like - I wasn't kidding about the 2A3-40 - its great and sort of affordable at £105 ea or £210 for a matched pair (HotRox) Easy to get an honest 6W output and a bit of effort gets 8W...

But value for money it isn't - the GU-50 is the value for money option here...

Of course I would go for PT15 or GK-71 but that just me

[Paul Barker]

So true, I will never forget my first disappointment with 300b and my massively unpopular public stance on it. I stood alone right then when it was flavour of the decade frenzy.

Also not forgotten the massively superior sound of the SV811-10. My point of ref at that time.

I came out about it in a private group “Triodeel” I think it was called. Stephie and Jono got it, the rest were following the 300b lemming March.

[JamesD]

I suppose the other thing to take account of is that the ear doesn't hear power increases linearly so 8W doesn't sound twice as loud as 4W - in fact it is only a little bit louder to our hearing so worrying about 6W or 8W from a power output perspective is unnecessary unless you are marketing the beasts...

811-10 and its big brother 572-10 are terrific in A1 - and even better with an A2 capable design More power to your elbow to get it finished!

[Cressy Snr]

There’s a KT120 model on the site, and the beauty of this calculator is that you can set your transformer impedance, fiddle with the current until the value of your existing cathode resistor comes up. This way you can experiment with substituting other valves into your existing design and be able make an estimate as to whether that valve it is going to sound good, crap, or whether it’ll blow your amp up.

I tried the KT120 triode model with my HT and fiddled the current until it came up with the cathode resistor value in my amp. It gave a 100mA op-point at -27V of bias, 3.8W output at 5K and the 2nd harmonic distortion wasn’t a lot worse than the KT77. The 3rd and 4th however were far lower than the KT77.

Substituting the KT120 into my amp, did indeed produce 100mA, but 26.5V of bias so well within any margin for error. I’ve not heard my KT120s sound anywhere near this good BTW and this is with puny 6SL7s driving them; so with a more suitable driver, it looks like a good candidate for a more powerful amp, without having to spend a great deal of dosh as the Hammonds can handle 138mA of bias. As it is in my amp at the moment, the KT120 is producing a nice musical sound and the tone is big and relaxed.

I’m not going to put up the pic of the curves because it will come out huge again, but the KT120 model appears accurate and as triode strapped pentodes go, it isn’t as bad as all that; certainly it looks better than the KT88.

The KT150 model looks very, very good indeed in triode mode and a nice 10W single-ended amp could be made with it, if your output transformers are any good. KT150s are a bit expensive for me though, which is a pity.

following his link to how the various curves are derived, it looks as if he’s using some sort SPICE modelling setup. Everything I’ve tried in my own amp, working from his models, has come out as predicted, so the models appear to be sound.

BTW, there’s an Osram PX25 in the list of valve models.

[Paul Barker]

“you can experiment with substituting other valves into your existing design and be able make an estimate as to whether that valve it is going to sound good, crap, or whether it’ll blow your amp up “

Some of us built after manually manipulating load lines.

“BTW, there’s an Osram PX25 in the list of valve models“

Cool, very few users though due to no true px25’s in production. Not much point studying Loadlines for unobtanium. Ok Steve and I have a couple originals. Steve more so. But we both know them well enough by their sound. Loadlines are for your preparations, sound completes the process! Then it’s over for tricks and following the crowd.

[JamesD]

Paul,

Have a look at the Emission Labs PX25 - they spent a lot of time trying to do it properly and it looks promising to me on paper...

of course it very expensive.... but cheaper than an original..

James

[steve s]

So true, I will never forget my first disappointment with 300b and my massively unpopular public stance on it. I stood alone right then when it was flavour of the decade frenzy.

Also not forgotten the massively superior sound of the SV811-10. My point of ref at that time.

I came out about it in a private group “Triodeel” I think it was called. Stephie and Jono got it, the rest were following the 300b lemming March.

The first time I came to your place in the late 90s Paul you still had a few 300bs on the shelf..
That was about the time I built my 1st px4 amp.. Great days.. I remember you coming to see it.. we compared it to a simply 845 amp (that I sold just after)

[steve s]

Paul,

Have a look at the Emission Labs PX25 - they spent a lot of time trying to do it properly and it looks promising to me on paper...

of course it very expensive.... but cheaper than an original..

James

I looked up emission labs.. some of the prices are way above some of the original valves. I was surprised to see those prices

[Paul Barker]

Well its good job they make them at least.

[ed]

Also not forgotten the massively superior sound of the SV811-10. My point of ref at that time.

just carry on bigging them up please.....

I've got quite a few to sell in the near future.