JLH 10 Watt class A amplifier

Geoff,

Sorry, my post was a little unclear. From the September 1996 Update pages:-

"As I commented, at the time, the design gave a somewhat lower distortion if the hFE of Tr1 was greater than that of Tr2. This caused the output circuit to act as an amplifier with an active collector load rather than an output emitter follower with an active emitter load"

Perhaps the 200 vs 160 hfe's was too great a difference and might have been self defeating and therefore not allow the 2N3055's to give their best? All I know is that to my humble (and of course subjective) ear'oles, the MJ's sound better in the amplifier. Bias current, however, must come into the sonic equation because curiously (and maybe someone has an explanation for this, if we really want to flog the subject to death) I tried the same substitution in a DoZ Headphone amplifier with my Senn 580's and they sounded worse, so I put the 2n3055's back in and the magic came back. The main difference is the headphone rail is half that of the power amp (23V vs 40V) and it's biased at 330mA per channel instead of 1.7A for the power amp.

PS Geoff, Thanks for the great website by the way.
 
That's right!!!

Thanks Geoff, the class-A website is indeed great.

Thanks to you and Rod Elliot I started this hobby and went on building a class-A power amp.

greetings, Thijs

PS did any of you ever see a Bedini 25/25 schematic.. an amp with such a lovely name much be a lovely amp.. I haven't found anything yet...
 
Sidney

No need to apologise, it's not the first time I've seen this comment in a forum post. I think the paragraph from the 1996 article that you quoted is a little misleading (perhaps due to some editorial interference) and that it should be read in conjunction with the table in the original 1696 article.

Another problem with the 1996 article is the suggestion that Tr1 be replaced with a Darlington to ensure that its gain is greater than that of Tr2. When I tried this five years or so ago, the only result was severe oscillation. I didn't bother to try and find a cure but just reverted to matched output transistors.

A possible explanation for your experience when trying the MJs in the headphone amp is the gain characteristics of the two devices. The performance of the JLH circuit is greatly influenced by the gain of the output transistors, as indicated in Table 2 in the original article. From measurements I have taken, the gain of a 2N3055 is around 280 at an Ic of 0.3A compared with 120 at 1.5A and less than 100 at 2A. Conversely, all the MJ/MJL devices that I have tested have had a gain that progressively decreases as the collector current is reduced from 2A to 0.1A

Thanks for your kind comments about my website, and thanks for yours as well Thijs.

Geoff
 
3055...

Dear Thijs,

2N3055's are very old design, yes. And they exhibit poor linearity, I cann't agree whith you. On my desk, the Motorola Power Data Book, I can see : Typical Dc current gain 150 @ 0.25 A , 80 @ 1 A , 50 @ 2 A. This causes important, audible harmonic distortion, at all frequencies.

The low current-gain-bandwith product (2 Mhz) causes very unpleasant intermodulation distortion at high frequencies.

Modern power transistors, specially designed for audio, are more expensive, but unquestionably more linear and fast, and give obvious inprovement of sound reproduction, specially with present sources, which can have noticeable energy in the high frequencies.

Regards, PL.
 
I'm also working on JLH96

I got my JLH96 project working some time in January this year. I went with newer output transistors right from the start. I use Sanken 2SC2837... Newer transitors indeed need steps to make the whole amp stable.

Additionally I have made one more bigger change until now. I have thrown out C4 (the el.cap. from feedback to ground). To compensate for the current that starts to flow through R6 I paralleled R1 and RV1 with 5030 ohm of resistance (4.7k + 330ohm). This left the DC offset pretty much where it was and it needed only minor adjustment..... I know, I know you guys think that it's not a good move as the DC offset will begin to drift with amp warming up. Well it does, but it is in order of 30-40mV total and I don't consider it to be of any concern.

The sound changed a lot more than I had expected. It was almost of the same magnitude as I experienced by removing electrolytics from output of Marantz CD67. One of the strenghts of JHL amp often described is how it plays piano music. Well that improved even more. The soundstage after the mod is even more open and a lot of air and freedom......

Anyway, I have some ideas to try out in coming months, one of them being a mission to get rid of Tr5 current source and C5. One idea being to substitute it with JFET current source....

I have had a plan to write all my experiences down and provide solid measurements and send it to Geoff to put on his site (we talked about that possibility with Geoff at the end of last year and he was willing to do that), so I don't want to ramble too much at the moment as I want it to be more tought through and I hope it will be helpful for other JLH owners/potential builders as a source of some ideas :)

Ergo

PS. I'm very happy with these amps already, but the tweakers hart demands improvements...
 
Dear P.Lacombe,

You are right. I read the D. Self book again last night and I was terribly wrong! Sorry about that. The 2N3055 does indeed have large beta drop in the 1.2A range. As I remmebered it was not so severe but after reading it again I must appologize. However the gain that is lost at higher current may not be there at all when using the mj15003 since it has lower gain to begin with. So the 2N3055 will provide more gain=feedback=less distortion at the most importend sound levels. I must say that I am very curiuos about the new transistors, but I am still not convinced: measurements and listeing trials will guve the final awnser.


Dear..ergo..

come on!!! show us the good stuff!!

greetings,
thijs
 
Dear P.Lacombe

The low current-gain-bandwith product make the amp stable without a dominating compensation cap. Since no turn on or off is needed the output device may not not to be faster?

Is there any point in using faster devices when it means instaling compensation components to slow things down?


I always see such cap around diff input stages which use voltage feedback. Sometime the input capacitance of MOSFETS takes care of that but isn't that the same priciple?

greetings,
thijs
 
3020

Eric

The 'Normal' input on the 3020 has steep (18dB/8ve) high and low pass filters with -3dB points at around 13Hz and 49kHz. This limits the flat passband to 20Hz - 30kHz. The 'Lab' input has no filtering and the -3dB points are about 6Hz and 138kHz.

JLH

Ergo

I agree with your desire to remove Tr5. However, you may be hard-pressed to use a JFet current source since, depending on the gain of your output transistors, the ccs will need to supply between 20mA and 60mA for the normal 2A quiescent current. A standard two transistor or LED/transistor ccs should be fine though (or you could even revert to a bootstrapped resistor :)

One piece of information that has yet to find its way on to my website is that replacing the 7815 and associated components with a css (as per the 'JLH for ESL' design) gives an audible improvement (smoother and more clarity). JLH himself used this arrangement on a 1975 high-power Class-B design with a similar front end and he also omitted the feedback capacitor. If you are interested I'll send you a copy of the schematic.

Geoff
 
Geoff, I'm very interested, please do send the copy if you'll find time...

As for cs with FETs I thought in lines of putting FETs in parallel (every FET gives about 5mA) and I could adjust the bias with adding/removing some. There are many who claim that in such place FET's give the best sounding solution. One of my friends has made a lot of experiments and also supports this finding.

At the moment there is one problem though. As the power supply gives me +/-32V lines it is a bit much for small signal FETs. I'm thinking of doing LC filtering in power supply that would give me about the same voltage as transformer secondaries.... which are 24V. This mod is waiting for cach to buy the chokes :)

***

tschrama, I will try to format some schematics and measurements to post here to give the idea of what I've done so far.

Ergo
 
The schematic that Argo has posted is the JLH 'Simple 30W integrated amplifier' published in Hi-Fi News Jan/Feb/Mar 1980 and updated in Dec 1980 to use MOSFET output devices in place of Darlingtons.

There are several errors in the schematic which I have previously pointed out to the site owner but which have not been corrected.

The speaker output should be taken from the junction of the 2000uF capacitor and the 4k7 resistor.

The supply rail voltage should be 55V to 80V depending on the power output required (30W to 50W).

As shown, the closed loop gain is set at 221 which is too high (particularly for a 25-30V supply rail). The original 30W design had a closed loop gain of 22 which would need to be increased to 29 for the 50W version to retain the same input sensitivity (0.7Vrms for full 8ohm output).

Geoff
 
Geoff said:
There are several errors in the schematic which I have previously pointed out to the site owner but which have not been corrected.

The speaker output should be taken from the junction of the 2000uF capacitor and the 4k7 resistor.

I was wondering about that speaker output also.

Site owner states that this amp is 30 Watt singelended class A. Is this true?

Argo
 
JLH 10 watt - does it make sense??

Hello...

I am new to this forum, but looked at the schematic for this 10 watt amplifier. I'm confused why anyone would want to build this one, does it make sense?

Don't get me wrong, I am 100% for building DIY amps of all types. I suppose that building anything is better than nothing, but...

The amp uses a single ended supply?? Why? This forces you to make it AC coupled on the output. Then we find a resistor to ground at the output?? Wha?

There are all sorts of really nice class A amps of modest power that are DC coupled in this power class (not that a 10 watt amp is something that is terribly practical) that have been published and many that have PCBs available.

I'd point any Euro based people to Erno Borbely's site, which has many designs of outstanding design and sonic quality.

My thinking is that if you have to take the time and money to build something from scratch, why not build something of significant usablility and quality?

Just my 2 cents... not trying to rain on the parade.
 
Keep on reading dear Bear.......


goodluck,
Thijs

well OK.. I will reply:

The single ended supply is easy to implement, only one regulator needed.
The output cap's also is an inherent speaker protection, no DC danger. This can offcource be modified but is a OK start. (Can you even hear DC sound?)
The resistor at the output is only there to prevent any charge building up with no speaker atached but can be left out.
10 Watts gives about 12dB headroom
No PCB is needed.. it's a very simple design with very good performance

There are many good thinks to say about the amp both from theoretical and practical experience. I do suggest you keep on reading.

Greetings,
Thijs
 
Argo

The amp you posted is most certainly not single-ended nor Class-A. It is a fairly standard push-pull Class-AB. Apart from the lack of an input differential, the main point of interest in the original (pre-MOSFET) circuit was the way the output Darlingtons had been used. By providing an additional driver and turning the Darlingtons upside down to form an output triplet, the usual problem of thermal coupling within the Darligton was avoided.

Bear

Thijs has offered some comments and I would add the following:

The JLH Class-A amp is perfectly useable and is of significant quality. It can be configured for a single supply rail or direct coupling and the power output can be set from 10W up to about 35- 40W (if the load impedance is known) depending on one's needs and depth of wallet.

It is by no means coincidence that this 30 odd year old design is still popular and is still being built in significant numbers around the world. The favourable comments received when these projects are completed bears testimony to the quality of the amp.

If I had more time, I could go on at length about the benefits of this circuit, but I will restrict myself to saying that one most desirable features is the inherent stability which obviates the need for additional compensation capacitors and the consequent loading on the VAS.

Yes, there are many other good designs around but, judging from the feedback I have received, the JLH will equal or better them in many subjectively important areas.

If I can find time later I will expand on these comments.

Geoff
 
  • Like
Reactions: 1 user