JLH 10 Watt class A amplifier

BD139 replace with 2SC5200

Hi,

Has any one see any problem for me to try the replacement Q4 for the high power updated version? It simulate ok with 20V supply rail and all output transistors the same 2SC5200 and Iq set at 4A. But I think in reality the rail will be a few volt higher if I use the CLC, or cap multiplier supply with the 300VA 22-0-22 transformer. If I remember correctly the voltage after the bridge rectifier with this transformer in my current JLH is around 28V before the LM388K circuit.

If the 2SC5200 is fine to replace BD139 then I will use five of them per channel.

Regards,
Chris
 
Re: Feedback Capacitor R8 (2k7)

ralf said:
Hi,

what does it matter to use 2k2 instead of 2k7?
(my Caddock -seller has only 2k2)

thanks,

Ralf


Hi,

The DC offset will increase a little, the amount of feedback is reduced.. you can try and listen for the difference between the two values with the normal brand resistors first and if you like the 2k2 sound then get the caddock ones, but make sure you reset the DC offset first after each replacement.
Chris


Sorry, shouldn't I say the amount of feeback is increased instead cos the values go from 2.7k lower to 2.2k so it allow higher feedback
 
Lm338 ps in a boosted version?

I' ve started my 30W JLH-2003 project.
PS will be regulated, with the LM338K.

However, I've been put somewhat in a doubt by one experience I had building my current 13 W J-amp which uses the 7815/2N3055 and 7915/MJ2955 ps.

During testing of my current amps power supply I somewhere short-circuited and after a minute or so I found the problem: all the current was flowing thru the Amp-meter. which indicated 10.2 or so. wires and connectors started to smoke etc etc . But there it was: 10 Amps available for more than .5 sec.

Doubt arises on the point if I should not "boost" the LM338 with another BJT to provide more than the 5A or 12A peak into the amp, like the 1996 ps does.
In the Application Notes of the 338 there is a 10A version, which uses an opamp to control 2 LM338's but that's no option for me.

As I lack the ability (and time) to design a circuit like this, I'll use the Lm388 circuit. But if anyone can help my to settle my doubts. : You're more than welcome.:bulb:

BTW: the cap-multiplier PS option is not one I desire.
 
JLH problem

Following a recent problem which came to my attention after my JLH rebuild, and in the hope that other constructors might be able to avoid a similar experience, I suggest that Q5/6/7/8, refering to fig 2 on http://www.gmweb.btinternet.co.uk/jlhupdate.htm should be MPSA56, not 2SA970 as shown.
The problem in question seems to be caused by possible oscillation (though this has yet to be confirmed) in the dc offset ccs and it is apparent through intermittent noises, sometimes also with hum, audible through the speakers. In my case the noises were usually evident within half a minute after switch-on (listening very close to the speaker), sometimes they could be heard across the room though at other times they were totally absent. Often they would sound like RF interference or hiss, touching the speaker terminals would often change the characteristics of the noise. When Q5/4 were changed to 2SA872 the problem became worse still. I know of another constructor who has apparently has similar problems using BC212s, so it seems the problem might not be limited to my circuit and 2SA970s/2SA872s. Installing MPSA56 transistors (as suggested to me by Geoff Moss) cured the problem.

Tim.
 
Hi TimA.

Maybe the problem is not the 2sa970 but more 2SC3421...

The miller capacitance cc is 15pF for 2sc3421 and 25pF for 2n1711.

try add a 10 - 22pF across basis and collector of 2sc3421 and use the 2sa970 again. Then i think the problem will disappear.

Even better use 100pF instead and the choice output transistors that will not oscillate is a lot more.

:clown:

Only a few thoughts from sonnya
 
Re: amount of feedback

ralf said:
Hi,

has someone tested the sound influence by changing the amount of feedback?

How does it sound with zero feedback?

greets,
Ralf


Ralf,

I do not think the JLH class A will work with zero feedback. I think it is one of those kind of design needs the feedback network in order for the amp to work.
As of trying different values and listen for it I have not yet.
But the value of this resistor is rather critical not only to the sound effects upon the amp but it is upmost critical to the stability of it, be careful, (IMHO) I am sure that zero value will destroy speakers...(experienced diyer please comment here)

Chris
 
Hi Chris,

I built the following amps in order:
1. A Hybrid Class B amp
2. Pass A40 (Class A)
3. Pass Aleph-3 (Class A)
4. Musical Fidelity A1 (Class A using exactly pcb layout)
5. JLH Updated version
6. JLH updated High Power(rail +-31v) Version using JLH1996 lq control
7. Hiraga 20W (Class A)

The sound of JLH updated out-perform amp 1,2,3,4 , extremely warm and details, relatively musical although a little bit darken, without affecting my listening interest(on classical musics), instruments appear clearly. I really touch the soul of music through JLH. The sound of amp 6 is quite a bit open and airly with taut bass, but the high is a little bit harsh. I do think it depends on personal taste which is subjective but I agreed with Tim Andrew findings post at Geoff's web page. Furthermore, the input capacitor affect the sound very much. I used only ordinary polyproprene cap as input, ordinary 1% metal film resistors(shortage of pocket). I wrote to share my little experience and let everyone know the schematics post is tangible. Thank you.

Rgds

John
 
Tiger vs A40

I have built a couple of Tigers and one A40, I think that the A40 was the pearly gate for me it opens my mind for high end sota designs... but what comes then? nothing as far as I can see, the Krell KSA series blow my mind, what is the f-ng secrets of their designs.

Back to planet earth the obscure fact is nothing, then just POWER a big POWER SUPPLY WILL DO IT.
 
JLH 'update' schematics

There are many references to Geoff's Class A amplifier site on this thread. Very unfortunately, current circumstances have meant that Geoff has had to close his superb site, at least for the time being. As recent references and discussion here have been with regard to the updated JLH circuit and the updated JLH for ESL circuit, I have attached them both here for future reference, or until Geoff's site is reopened.

Tim.
 

Attachments

  • jlhupd~2.jpg
    jlhupd~2.jpg
    32.5 KB · Views: 3,468
Geoff's JLH amp site

As my current project is the 2003 JLH -ESL- version, one design change is not noted in the diagrams of TimA.

I've incorporated this change into this diagram:

An externally hosted image should be here but it was not working when we last tested it.


Its the C5 which is connected to a split R11/R10 . It is there to reduce
hum from unregulated power supplies , but I do use it in a project with a regulated ps: it makes good sense.
 
jccliu said:
Hi Chris,

I built the following amps in order:
1. A Hybrid Class B amp
2. Pass A40 (Class A)
3. Pass Aleph-3 (Class A)
4. Musical Fidelity A1 (Class A using exactly pcb layout)
5. JLH Updated version
6. JLH updated High Power(rail +-31v) Version using JLH1996 lq control
7. Hiraga 20W (Class A)

The sound of JLH updated out-perform amp 1,2,3,4 , extremely warm and details, relatively musical although a little bit darken, without affecting my listening interest(on classical musics), instruments appear clearly. I really touch the soul of music through JLH. The sound of amp 6 is quite a bit open and airly with taut bass, but the high is a little bit harsh. I do think it depends on personal taste which is subjective but I agreed with Tim Andrew findings post at Geoff's web page. Furthermore, the input capacitor affect the sound very much. I used only ordinary polyproprene cap as input, ordinary 1% metal film resistors(shortage of pocket). I wrote to share my little experience and let everyone know the schematics post is tangible. Thank you.

Rgds

John


Hi John,
Thanks for this feedback,
Here's mine:

In 2001 I built a dual mono JLH-1996 with basic components (1% mf resistors, ST 2N3055's, "no audio grade C1" , basic capacitors, BC056 caps as reservoir etc etc ).
After completion I was stunned. I ended up buying a secondhand NAD pre-amp which feeds the input from my Denon set CD player/tuner and my Technics SL1900 to my project #1.

My J-amp filled a "nag" I had for "short on low". I was blaming my small JM-Lab speakers, but I should have blamed Denon at least.

Now I'm building the 2003-ESL version with mostly audio-grade stuff: Jensen 0.47uF (alu foil ok, as 2 copper foil Jensen caps would add another 100 euro ) for C1, HolCo resistors, soft-recovery diodes for the bridges; BC154 as reservoir caps; Elna-Silmic and Panasonic FC's for other capacitors etc etc More heatsinking etc.
Design is to run rail voltage at 28V with a Iq of 3.2 A.

I can't wait to hear the result. But I'll let you know....
 
Re: Tiger vs A40

kamskoma said:
I have built a couple of Tigers and one A40, I think that the A40 was the pearly gate for me it opens my mind for high end sota designs... but what comes then? nothing as far as I can see, the Krell KSA series blow my mind, what is the f-ng secrets of their designs.

Back to planet earth the obscure fact is nothing, then just POWER a big POWER SUPPLY WILL DO IT.

Hi Kasmkoma,

PowerSupply is the key factor on building large amps.

Any designer will be able to design a regulated powersupply which can produce 40 A at 80 V.

Amp designers know how to design output stages which will deliver this amount of power.

The cost however to have a regulated powersupply which can support this output stage (40A at 80V) are tremendous.

As the current which has to be delivered at the terminals flows through the PS at best 1 PS active-device will be able to support 2 AMP active devices.
( power dissipation is lower in the PS section)
So you have an Aleph X-amp with 32 output FET's on a rail.
There is a unit with 16 devices needed for a regulated PS. )

They will not do that.

The alternative: building unregulated, but big-enough is what even Pass does.

I'll stick to "regulated PS" :bigeyes: