JLH Class-A, component help

[dominic]

Hello All,

This is my first post so please be gentle...

I've just built a JLH Class A (with a few modifications...mainly using what I had to hand). It's currently running a pair of 50+ year old GEC metal cone speakers at a few watts each max from an old UPS battery I had lying around...

I'm not too concerned about much more power, the speakers would probably melt and they're pretty loud as it is.

It sounds absolutely fantastic, far better than I thought and after trying out on my little kiddies breadboard kit and frying a few transistor, I've built it into a aluminium box which seems to be good enough at dissipating the heat.

But I wondered, before I go off and build a power supply for it what comments the experts would have on the components I've used and whether I should go out and buy "proper" ones before going any further.

For the output drivers I've used a couple of BD711s, these are specced at 100V/2A/75W they've all got hfe of about 130 not matched but with lower hfe's at the top of the circuit.

The drivers are BC337-16 and BC307As.

I know these aren't high grade components but how much difference would better spec parts make.

Has anybody tried using good quality power Darlingtons for the output pair.

I'm sorry if these all seem like daft questions but it's such a simple circuit and sounds brilliant as it is don't want to tamper with it unless it will make a bigish difference

TIA

Dom

[Geoff]

Hi Dom

Welcome to the forum and to the happy band of JLH builders.

It is difficult to comment on your choice of transistors without knowing your present/intended supply rail voltage(s) and quiescent current (and, ideally, the JLH version you have built). Details of your required power output and speaker impedance would also be useful.

I haven't used a pair of darlingtons in the output, but I have tried a single darlington for Q1, as suggested by JLH in the 1996 article, and the result was severe oscillation. I didn't attempt to find the cause of this (it could have been due to layout/wiring deficiencies) but simply reverted to a standard BJT.

If you haven't already been there, check out my website via the button below.

Geoff

[hlyytine]

Hello Geoff,

I haven't used a pair of darlingtons in the output, but I have tried a single darlington for Q1, as suggested by JLH in the 1996 article, and the result was severe oscillation. I didn't attempt to find the cause of this (it could have been due to layout/wiring deficiencies) but simply reverted to a standard BJT.[/B]

I have not experimented with darlington or CFP outputs either, but I have tried CFP arrangement at the input. Another attempt was to use a small-signal NPN transistor to drive the phase splitter. The collector of this helper transistor was connected to ground instead of the collector of the phase splitter in an attempt to isolate the capacitance of the splitter's collector and base from the input stage. It worked somehow, at least with a small compensation cap. The slew rate was affected, though. Very authorative bass and hypnotizing middle and treble, I really liked the sound. However, seeing the ringing on the scope bothered me and there we go...

So, while it did not openly oscillate, it had lots of ringing. I removed those extra components Monday, but it began to oscillate heavily. Now it seems I forgot to mention I increased feedback by changing the 2.7k resistor to 1.621k. Swapped the original 2.7k back in and the all-out oscillation turned into severe ringing. Now I tried what tschrama suggested in an another thread and an another context, replaced the 220R resistor from the feedback node to ground with a 390R resistor. In order to keep the amount of feedback the same, I had to increase the 2.7k resistor to 4.7k. After that the ringing almost disappeared (only two beats barely visible on the scope at full drive).

I cannot comment on the impact to the sound quality yet, since now the amps take their power directly from a capacitor bank. Earlier they had a dedicated capacitance multiplier per channel (a little modified, the positive rail uses a MJ2955 so it's a CFP connection like in the negative rail).

[dominic]

Cheers,

The speakers I'm using are pretty efficient, I'm told by the previous owner they're 2.5 Watts max each and he reckons about 15ohms impedance. They are very loud running off a 12V UPS battery at the moment. So I doubt I'll go much higher than 15-20 volts single rail supply for a wee bit more head room maybe as high as 25V.

I have to say they sound absolutely fantastic, I'm told that the speakers when new cost more than a brand new jaguar, although he used to have eight of them!

I don't mind having the output capacitively coupled - though I will probably at some point try and get some bigger/better quality output caps. The signal capacitors for input decoupling and signal path are the most expensive parts so far!

Looking at things on a scope the whole thing seems to be running very well - not bad for a fast stab - no ringing and square waves come through completely unscathed to the coupling capacitors with the load attached.

Another question I forgot to ask is how do I calculate the optimum quiescent current, running at 550mA per channel at the moment, going higher doesn't seem to affect the sound but don't want to go much higher to avoid burnt fingers!

The question about the parts was more about any possible impairment due to the cheap nature of the parts, though I guess I'd struggle to tell the difference until I fork out for a new CD player!

Cheers

[Geoff]

Originally posted by dominic
Another question I forgot to ask is how do I calculate the optimum quiescent current, running at 550mA per channel at the moment, going higher doesn't seem to affect the sound but don't want to go much higher to avoid burnt fingers!

For a single supply rail, the peak output voltage will be half the supply rail voltage less an allowance for losses (say 1 to 2V). The peak output current will depend on the minimum impedance of your speakers. As you a using a single full-range driver this can be determined by measuring the voice coil resistance (which will equal the minimum impedance as near as dammit).

The peak output current will be the peak output voltage divided by the minimum speaker impedance. The optimum quiescent current for the JLH is around 0.75 times the peak output current.

Geoff

[dominic]

Cheers Geoff

Thanks, when I measure it with a multimeter it's showing 3.75 ohms, not the 15 I was expecting, however it's not a single driver unit it's got a screw in "presence" driver (tweeter) quite fancy!

(6-1)/3.75 = 1amp I guess, so I should turn it up, think I might steal some heatsink fins from somewhere else before I do that!

Dom