Go Back   Home > Forums > >
Home Forums Rules Articles diyAudio Store Blogs Gallery Wiki Register Donations FAQ Calendar Search Today's Posts Mark Forums Read

Solid State Talk all about solid state amplification.

sublimed JLH1969
sublimed JLH1969
Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 4th November 2018, 11:59 AM   #1
kokoriantz is offline kokoriantz  Lebanon
diyAudio Member
 
Join Date: Dec 2015
Location: south east asia
Default sublimed JLH1969

The purpose of this thread is not to present an Nth version of JLH amplifier, but to apply to the JLH amplifier another technique of amplifier theory.
My theory is that, an amplifier should have all its output currents already circulating in its veins , after which ,it is the amplifier and not the PSU who is providing the power to the output, particularly at the transients . By this, the power supply delivers a continuous dc current , which means ,switching PSU can be used .I tried out this theory successfully on small scale using TDA1552Q in bridged class A , with extraordinary result . (see The Moon for six pence). the Moon for six pence

Instead of using two amps bridged I wanted to experiment now a single amp, bridged with a paire of capacitors . It also results with a single power supply configuration ,and also dc current only from the supply. Now ,what amp ,already well functioning and available in kit at affordable price, I would think of to experiment upon ? The JLH of course . So I bought a Chinese JLH1969 for 7.5$ in kit to modify .

Once I applied the basic modification , I was expecting enhancement in high or mid frequencies as occurred precedently ,but to my great surprise I could hear tube- amp quality bass sound . With further adjustments I obtained an unprecedented low frequency performance . With all technologies considered ,I have never heard such bass , as if I added a subwoofer to the system . Now this amp is going to be my refrence amp on this domain. Using another reference amp only in high frequency domaine , I succeeded to bring the JLH up to its 95% . Now the JLH1969 got sublimed ,adding both a subwoofer and a super tweeter to the system, delivering the widest spectrum of tonality I have ever heard.

I would feel myself guilty to keep it to myself such an extraordinary sounding amplifier ,which is so easy to built and so affordable that I hope DIY will be greatly satisfied .

https://www.youtube.com/watch?v=9IlT2yeOJ0g. My high frequency generator.
https://www.youtube.com/watch?v=1PdcmFvZvuI&t=570s. from 9:40 to13:50

These are the main sound sources I used to tune this amp for perfect reproduction .

Try to be pleased by these two musical pieces on any other amp than this one .
Kokoriantz.
Attached Images
File Type: jpg Sublimed JLH 1969.JPG (117.1 KB, 2283 views)
  Reply With Quote
Old 4th November 2018, 12:03 PM   #2
huggygood is offline huggygood  France
diyAudio Member
 
Join Date: Mar 2011
Location: france (bretagne)
my next assembly on veroboard
  Reply With Quote
Old 4th November 2018, 12:05 PM   #3
kokoriantz is offline kokoriantz  Lebanon
diyAudio Member
 
Join Date: Dec 2015
Location: south east asia
Default technical details

The main reason of this circuit as mentioned earlier, is to function with a constant current power supply .
The driver orders the outputs carrying the current A , increase in one and decrease in the other by i amps . The PS seeing the one output asking less current sparks the i current in excess thru its opposite capacitor by inertia (induction) trying to join its other pole. Meanwhile the second output which was receiving the current A from the first now gets i amps less , which in turn due to its inertia will pump it from its opposite capacitor via the load . The PS i current now cannot join the other pole by the second capacitor carrying an opposite transient current ,but must flow via the load and the second output , which has received order to acquire the extra i amps . By this the load sees both transient currents pass thru it and this on both phases . This yields a superior dynamic sound than single ended tube amps where only one phase , upon the collapse of the electromagnet, that such phenomena occurs.

The PS now sees the amplifier demanding a continuous current A ,but there is also the bootstrap capacitor . When the upper output voltage increases, the capacitor discharges thru the upper transistor as an autonomous circuit , but when the output decreases, the capacitor is charged by the lower transistor from the PS. this provokes a dissymmetric ac current ,but only 1/16 of the output current . A switching PSU can be used successfully providing 27v 1.25A . A 50watt shall be sufficient.
The load now is referenced to a floating point , whereas the input is referenced to 0v . It is natural that the floating center point becomes now the reference common ground and the whole circuit instead is floating. This why each channel needs its own PS. The reactance of the output capacitors do not take part any more of the feedback . It also allows the coupling capacitor to fly away. Now the input transistor gets its current from the devider emitter resistor via the load ,in order to keep it supplied unloaded , a pull up resistor is necessary . How much is the current that the input transistor needs ? A miserable 200u amp. The transistor is working in starvation regime! By humanity ,I decided to redesign this amp properly .

The designer was obliged to run it in starved mode ,as pulling more current by decreasing the collector resistor would decrease the open loop gain . So I added a CCS made by a paire of bc337 (2n2222 could have been used in Oxford) in current mirror, but of how much current? Auditively , up to 3ma there is enhancement, so let it be 3.5ma by 8.2k pull up resistor . This doubled the input's gain . Now I can reduce the collector resistor , Additively down to 4.7k gives considerable quality gain but later on simulator, it showed 3.3k gives the lowest distortion ,so let it be 3.3k . The transistor now has 4ma bias current and 3.9k pull up resistor can bring to 13.5v the output capacitors without load.

The base of the lower output transistor with its 15 ohm impedance sees the emitter of the driver now as 22ohm . The 2n3055 has particular linearity in Vbe/Ic transfer function and it pleases to be driven in voltage rather than current ,but what is this 2.2k resistor's function ? Normally it pulls extra current from the driver so that when the lower transistor reaches to cutoff ,it remains comfortably biased and it's not the case here. Additively again I decreased the value to 1k for better sound ,below is the same . On simulator it showed slight decrease in phase shift above 5 khz . The 2n3055 is of old type transistors that has along Ft a Fhfe of >10khz according Siemens datasheet. Motorola ones are may be 20khz. Normally a 3 stage amplifier requires a miller capacitor acting as dominant pole to achieve stability . With 2n3055 the dominant pole is inherent all naturally .The impedance curve which reflects the invers function of the open loop response shows clearly the 20khz dominant pole. It is possible that the 1k resistor is increasing the frequencies of other poles linked to the transistor as it shunts the b-e and c-b capacitors ,and leaves the dominant pole act alone, while the emitter follower at transients goes high impedance due to its b-e capacitor . In other words, if 2.2k was adequate for MJ480 ,1k is for 2n3055.

The circuit, Chinese JLH1969, has placed 220 ohm 1/4w(dissipating 1/2w) pull up resistor and adjustable resistor to be 68 ohms for 1.2A ,along bootstrap capacitor of 220uf. Across 220 ohm measured 10v,. that is a dc current of 45.5 ma to make the two outputs conduct 1.2A each . This gives an average beta of 50 . The 68 ohms shunts the b-e along with the bootstrap reducing its dynamic hfe by 20% . To higher the shunting resistor I swapped the resistors , 1w 220 ohm shunt and 2w 68 for pull up , the bootstrap capacitor to be 1500uf. Now the sound became sublime . Using the simulator it showed that the best value for the shunt is 250 ohm , this means the pull up should be unacceptable 39 ohms .

The transistor with the highest beta than , should be placed on the upper side ,to compensate the 220 ohm shunting of 7%. The 68 ohm receiving the same output ac voltage can dissipate 1.2w .

The JLH characteristics are measured by mj480 with beta of 140 . The open loop gain using 2n3055 of beta 50 requires increase in gain at the input stage ,as now it is possible . By how much ? To bring the output impedance to 0.16 ohm it needed to be doubled. The emitter resistor of the input stage 220 ohm is replaced by 100 ohm to double the open loop gain and the feedback resistor becomes 1.2k . Now the amplifier complies with the JLH specifications .

This amplifier as the original, requires also a roll-off capacitor of 220pf shunting the 1.2k feedback resistor, after which the frequency response becomes Bessel 570khz instead of Butterworth 100khz for the original . JLH advises not to use any , I leave it up to you to decide eider perfect square wave response or high slew rate .

The input bias voltage I transformed it into current bias by 470k . This way it does not require a supply filtering , the input capacitor with 470k does filter , but not on sub frequencies where it introduces a feedback from the floating supply . This gives a second order high cut function with its resonance and Q . The input R3C1 adjust these parameters . Similar function results in amplifiers using servo for automatic offset adjust. The input capacitor C1 must be more than 10uf to ensure low distortion at low frequencies . An input volume of 2k is possible .

This amp wakes up and sleeps politely .
See also " modification underway " post.

Kokoriantz
Attached Images
File Type: jpg Sublimed JLH 1969.JPG (117.1 KB, 2305 views)
File Type: jpg graph4.JPG (62.2 KB, 2149 views)
  Reply With Quote
Old 4th November 2018, 12:08 PM   #4
kokoriantz is offline kokoriantz  Lebanon
diyAudio Member
 
Join Date: Dec 2015
Location: south east asia
Default modification underway

The bootstrap circuit as seen has two problems to resolve .
The shunt resistor needs to be 250 ohm for best frequency responce of the upper output transistor.
The pull up resistor is loading the output too heavily pumping 12% from its power and polluting the PS .
The remedy I will try out in compliance with 1969 technology , is to supply the bias current by a choke . 1H/50ma can be sufficient , but its resistance should be less than 30 ohm . The choke must be small enough to be installed upon the heatsink adapter mounted by the same screws joining the heatsink .
I will go DIY mode and wind it myself . I'll report about it in few months .
Attached Images
File Type: jpg modif.JPG (24.3 KB, 2449 views)
  Reply With Quote
Old 4th November 2018, 12:14 PM   #5
kokoriantz is offline kokoriantz  Lebanon
diyAudio Member
 
Join Date: Dec 2015
Location: south east asia
Default construction

To built this amp you need to buy in kit the Chinese JLH1969 .
https://www.aliexpress.com/item/2017new-1pcs-diy-kits-JLH-1969-class-A-amplifier-Board-high-quality-PCB-Assembled-MOT-2N3055/32800895221.html?spm=a2g0s.9042311.0.0.3da24c4d1y3 S5A there is also on ebay .
The kit has genuine NOS transistors with best 2n3055 motorola, dark grey steal feet (hard to cut). Beside the transistors and accessories you need the 8.2k and 100k adj , (led + 10k) from the kit.


You need to buy / channel.
5 x bc337/(25 or 40)
100 ohm, 1k, 1.2k, 3.3k, 470k, 3.9k/0.5W, 220ohm/1W, 68ohm/2W
10uF/6.3v audio ( or 100uf from kit) , 1500uF/25v, 2x10000uF/25v Low ESR
small universal fiberglass board.
6 screw isolatorfor to-03 /220 , 7 ring terminals for 2.5mm screw , copper wire 1mm 40cm.
3 screw 2.5mmx20mm + nuts, 2xscrew 3mm and washers for heatsink.
Heatsink and 24v/2A minimum switching power supply .

Before mounting the heatsink adapter upon the pcb , you must cover with an adhesive tape all the holes of the output transistors and the new two holes drilled and chamfered . One pair of isolators get their cylindrical parts cut to isolate the edge screws from the pcb . The other four need the circles trimed to be whole cylinders to insert in the heatsink holes of collector screws .

The pcb needs to have 3 tracks cut . The bootstrap capacitor must not link the collector but the emitter of the other transistor in star point . The 100k adjust have the central and an edge link cut as well as the edge to R3 . The central links instead R3 and the edge the 0v (precedent ground) . All the component refrences on the board are correspondence with the schematic. R8 is replaced with a strap bearing 27v to feed the pull up resistor of the CCS. R2&R10 are free.

One of the 2n3055 bears a handwritten number (higher beta). It must be installed near the output capacitors.

The heatsink used is 8inches wide 1 1/2 inches high and about 4cm thick . Sited on table rises at ambient T 30C to 66C.

You always get a matched pair from the same lot of 5 pieces . To match, measure without touching the transistors with a digital multimeter on diode position . The red probe shorts b+c , The voltage drop should be equal.

https://www.aliexpress.com/item/Seven-Neon-Free-shipping-high-quality-ultral-thin-AC100-240V-DC-24V-4A-96W-switching/32799434842.html?spm=a2g0s.9042311.0.0.27424c4d18Y yEG
Attached Images
File Type: jpg DSC_0575.jpg (606.6 KB, 1958 views)
File Type: jpg DSC_0577.jpg (695.1 KB, 621 views)
File Type: jpg DSC_0578.jpg (831.1 KB, 640 views)
File Type: jpg DSC_0579.jpg (653.5 KB, 489 views)
  Reply With Quote
Old 4th November 2018, 12:19 PM   #6
huggygood is offline huggygood  France
diyAudio Member
 
Join Date: Mar 2011
Location: france (bretagne)
I already have this model all mounted, I have more than the modified
  Reply With Quote
Old 4th November 2018, 12:48 PM   #7
anti is offline anti  Slovakia
diyAudio Member
 
Join Date: Jun 2008
Funny thing is that you mention 1K being the optimal value for the 3055 b-e resistor. I empirically arrived at 1K2 (with TIP3055 or small MJE3055). I even mentioned that in the main jlh thread a couple days ago. I'm also seeing some other similarities in the bootstrap (not the choke).
  Reply With Quote
Old 4th November 2018, 01:22 PM   #8
hsiehinjesus is offline hsiehinjesus  Taiwan
diyAudio Member
 
Join Date: Jan 2006
Location: Taiwan
Great design
  Reply With Quote
Old 4th November 2018, 05:30 PM   #9
huggygood is offline huggygood  France
diyAudio Member
 
Join Date: Mar 2011
Location: france (bretagne)
I can not wait to try this one.
I have been blocking for a few months on tube electronics,
it feels good to return to known land
  Reply With Quote
Old 5th November 2018, 03:13 AM   #10
nigelwright7557 is offline nigelwright7557  United Kingdom
diyAudio Member
 
nigelwright7557's Avatar
 
Join Date: Apr 2008
Location: Carlisle, England
The good old 2n3055.
Farnell still sell them.

I first used them in a 225WRMS Maplin amplifier in 1980.
The amp worked well for my mobile disco for years without issues.
__________________
2020 versions of PCBCAD51 and PCBCAD360 out now >>> https://www.murtonpikesystems.co.uk
  Reply With Quote

Reply


sublimed JLH1969Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump

Similar Threads
Thread Thread Starter Forum Replies Last Post
JLH1969 CHINESE onuhannes Solid State 2 10th November 2015 01:27 PM
jjh amp like jlh1969 mos lurgee Solid State 0 2nd March 2013 03:37 PM
JC2012Doomday - The High Efficiency Version of JLH1969 Class-A jazzclassics Solid State 35 23rd September 2012 05:06 PM


New To Site? Need Help?

All times are GMT. The time now is 11:37 AM.


Search Engine Optimisation provided by DragonByte SEO (Pro) - vBulletin Mods & Addons Copyright © 2019 DragonByte Technologies Ltd.
Resources saved on this page: MySQL 15.00%
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2019 DragonByte Technologies Ltd.
Copyright ©1999-2019 diyAudio
Wiki