I do indeed appreciate friendly atmosphere in this conversation. 
With your help I have managed to successifully build a 2003 version. Thanks OldDIY for your advice regarding the layout of the heatsink. The result is shown on the photo below. I have already prepared the amplifier for mounting into a DIY chasis, from rock-solid 3mm aluminium plates on a TIG welded AISI 304 stainless steel frame.
Also, special thanks to John-Ellis for instructions for biasing and drift adjustments in post #6994. That was indeed helpful.
I would like just to add that during the safety bulb test there is no sound in speakers because this amplifier needs more power to play. Just ignore the silence if there is no smoke and no sound. The amplifier will start singing as soon you feed it with enough callories, ... I mean Amperes. This guy runs indeed hot.
With your help I have managed to successifully build a 2003 version. Thanks OldDIY for your advice regarding the layout of the heatsink.
The result is shown on the photo below. I have already prepared the amplifier for mounting into a DIY chasis, from rock-solid 3mm aluminium plates on a TIG welded AISI 304 stainless steel frame.Also, special thanks to John-Ellis for instructions for biasing and drift adjustments in post #6994. That was indeed helpful.
I would like just to add that during the safety bulb test there is no sound
in speakers because this amplifier needs more power to play. Just ignore the silence if there is no smoke and no sound. The amplifier will start singing as soon you feed it with enough callories, ... I mean Amperes. This guy runs indeed hot. Attachments
The problem is that the input filter capacitor (C3 in #7478 that OldDIY mentions) and the input capacitor will try to turn the pnp on, but the feedback decoupling capacitor keeps it off. During that moment the pnp is off, the driver transistor is off, and that allows the upper resistors to turn the upper MOSFET on. I don't see that changing the capacitor values will have much effect, and a slugged PSU (capacitance multiplier) will really be needed.
@Kozard - I don't think your "ripple magnifier" will be too popular
Or maybe the bootstrap resistors could be divided into three? A small resistor at the top decoupled to ground with a big C shold stop that quick build up. BUt the resistor will need to be smallish (22 ohms or so?) so the decoupling cap will need to be big but that might be the simplest option???
And the capacitance multiplier/ ripple filter could be applied to the bootstrap/drive stage and not have to have a high current transistor?
@Kozard - I don't think your "ripple magnifier" will be too popular
Or maybe the bootstrap resistors could be divided into three? A small resistor at the top decoupled to ground with a big C shold stop that quick build up. BUt the resistor will need to be smallish (22 ohms or so?) so the decoupling cap will need to be big but that might be the simplest option???
And the capacitance multiplier/ ripple filter could be applied to the bootstrap/drive stage and not have to have a high current transistor?
Last edited:
Thanks everyone for your time and your attention!
Now that I soldered back the 4.7uF, the issue seems resolved.
I get a 26V peek to peek for only 2uS, which is perfectly safe to me!
Trying larger bootstrap and feedback capacitors is on the way and I'll keep an eye for possible improvement.
Is it safe for this amplifier to have the output shorted?
If it is, then this is an excellent solution!
This is a valid solution. But I wouldn't trust others that may operate the amplifier will follow the procedure.
I didn't know that a capacitance multiplier can help with this issue. The more I learn, the less I understand that I know I'll build a circuit soon to test it!
That's how my amplifier is. But!... at the assembly I exchanged the included low quality capacitors with better ones and thoughtlessly increased the C1 to 100uF.
Now that I soldered back the 4.7uF, the issue seems resolved.
I get a 26V peek to peek for only 2uS, which is perfectly safe to me!
Trying larger bootstrap and feedback capacitors is on the way and I'll keep an eye for possible improvement.
Is it safe for this amplifier to have the output shorted?
If it is, then this is an excellent solution!
This is a valid solution. But I wouldn't trust others that may operate the amplifier will follow the procedure.
I didn't know that a capacitance multiplier can help with this issue. The more I learn, the less I understand that I know
I'll build a circuit soon to test it!JLH 1969M Sen Yuan
For reference and for anyone that may find it useful, I tried to draw the schematic of this version, hopefully without many errors.
So far I'm very happy with how gracefully this little circuit drives my difficult speakers.
Thanks again for the tips and the help!
For reference and for anyone that may find it useful, I tried to draw the schematic of this version, hopefully without many errors.
So far I'm very happy with how gracefully this little circuit drives my difficult speakers.
Thanks again for the tips and the help!
Attachments
Last edited:
Ι've climbed up to 500mA per channel before the small temporary heatsink became too hot, circa 80C. Iq rises even more ~15% there. Higher Iq makes perceived sound more solid and clear.
A small chassis with larger heatsinks + kapton tape insulation + original mosfets are on the way where I plan to balance the Iq and the thermal dissipation within reason. Till then Iq is 80mA because it is on all day. Does not sound bad at all and I cannot stop listening.
Ideally, such an amplifier would have a front panel switch to select two or more predefined Iq. This could also help to keep the heating bill low during the winter
Any ideas on rating the thermal stability?
Thank you!
Last edited:
Install a TO126 transistor between the resistor in the driver emitter and ground. Close B-E (Diode switching). Install on MOSFET. Decrease current (or resistor value).
Didn't try
Thank you, this sound very interesting. I have to admit that I didn't understand the exact implementation. Is there a schematic, preferably with values?
Well, for a week now I have been listening to jlh 1969 (the PCB has been on the shelf for a couple of years - I finally soldered - I wanted to try the famous JLH a long time ago) - I have to say that it sounds respectable - good bass and very enjoyable mid-range and top. Very good for such a simple circuit, I agree with PMA.
For now its a "naked" construction - maybe in the spring I will make a wooden "coat" for it and i need a PCBs for PSU...and maybe bigger transformer or two of them.
PSU 3x4700 uf (actually 5200uf measured) 1R and another 3x4700uf for each channel. Output 4700Uf, input 2uf2. On the board there is an additional 2200Uf and 2200uf filter for the input transistor. 23.5V power supply, MAX. 7.7W for 8 and 4 R loads. Quiescent current 1.1A (power supply max. 1.25 A - so a little reserve is left + transformer will heat up less).
At input BC 560, VA / phase splitter-D139 (with hfe 140 approx.) - at output BD911 also Hfe around 140 - I selected four out of ten with the largest and equal hfe. Admittedly they are pretty accurate - half around 70 and other half 140.
p.s. Remember that on these ebay PCBs BC560 must be turned upside down than indicated on the board.
For now its a "naked" construction - maybe in the spring I will make a wooden "coat" for it and i need a PCBs for PSU...and maybe bigger transformer or two of them.
PSU 3x4700 uf (actually 5200uf measured) 1R and another 3x4700uf for each channel. Output 4700Uf, input 2uf2. On the board there is an additional 2200Uf and 2200uf filter for the input transistor. 23.5V power supply, MAX. 7.7W for 8 and 4 R loads. Quiescent current 1.1A (power supply max. 1.25 A - so a little reserve is left + transformer will heat up less).
At input BC 560, VA / phase splitter-D139 (with hfe 140 approx.) - at output BD911 also Hfe around 140 - I selected four out of ten with the largest and equal hfe. Admittedly they are pretty accurate - half around 70 and other half 140.
p.s. Remember that on these ebay PCBs BC560 must be turned upside down than indicated on the board.
Attachments
JLH 1969 to direct drive ribbon speaker
Hello Forum,
I built the JLH 1969 original version on stripboard (BC560c, BD139-16, MJ15003). Works so far on small bookshelf speakers.
My idea is to adapt the original design to direct drive a ribbon speaker. This one will be operated from 200-20000 Hz (laminated for higher mechanical resistance) and has a pretty stable impendance of about 1 Ohm.
PSU are SMPSs from Lenovo notebooks and deliver 20V 4A. I plan to upgrade those with thermistors + external buffer of some 10.000 s uF for a more stable current supply.
In the original article from 1969 (TCAAS) there is a summary for different load impendances from 15 to 3 Ohms. To start i thought to halve the R1 and R2 values and double C1 and C2 and measure distortion, then adapt step by step.
I didn't find a reference for use of this design with such a low impendance. What else to consider? Opinions? Hints? ..Thanks in advance!
Hello Forum,
I built the JLH 1969 original version on stripboard (BC560c, BD139-16, MJ15003). Works so far on small bookshelf speakers.
My idea is to adapt the original design to direct drive a ribbon speaker. This one will be operated from 200-20000 Hz (laminated for higher mechanical resistance) and has a pretty stable impendance of about 1 Ohm.
PSU are SMPSs from Lenovo notebooks and deliver 20V 4A. I plan to upgrade those with thermistors + external buffer of some 10.000 s uF for a more stable current supply.
In the original article from 1969 (TCAAS) there is a summary for different load impendances from 15 to 3 Ohms. To start i thought to halve the R1 and R2 values and double C1 and C2 and measure distortion, then adapt step by step.
I didn't find a reference for use of this design with such a low impendance. What else to consider? Opinions? Hints? ..Thanks in advance!
Attachments
The table isn't a measure of the amplifier's output impedance or damping factor. All it tells you is how to adapt the design to safely drive loud speakers of those differing impedances and still get the most from your basic amp.
Unless you plan to use low impedance loudspeakers, you will only be reducing the dynamic range and power available to your speakers, assuming they are more like 8R than 3R impedance. The table suggests resistor and other component values, supply and input voltages that allow higher bias and output currents which suit low impedance speakers but at the cost of lower output voltage. This means means less power for high impedance speakers.
So, its a table of compromises to support different speaker impedances. The reason being the limited power output capability of the single pair of power transistors operating in a compact, economical class A design amplifier. It certainly does not substantially change the amplifier's output impedance.
If you just want more power, think of a bigger, multi-transistor and driver stage class A amplifier design. In my experience, you may well be able to increase the JLH's power just by increasing the number of output transistors, heatsink size, transformer etc. but something will almost certainly be lost from the sound quality.
Unless you plan to use low impedance loudspeakers, you will only be reducing the dynamic range and power available to your speakers, assuming they are more like 8R than 3R impedance. The table suggests resistor and other component values, supply and input voltages that allow higher bias and output currents which suit low impedance speakers but at the cost of lower output voltage. This means means less power for high impedance speakers.
So, its a table of compromises to support different speaker impedances. The reason being the limited power output capability of the single pair of power transistors operating in a compact, economical class A design amplifier. It certainly does not substantially change the amplifier's output impedance.
If you just want more power, think of a bigger, multi-transistor and driver stage class A amplifier design. In my experience, you may well be able to increase the JLH's power just by increasing the number of output transistors, heatsink size, transformer etc. but something will almost certainly be lost from the sound quality.
Connect the output transistors in parallel (as in JLH2003) to maintain the allowable heat dissipation power.
Increase the driver current. Double the quiescent current compared to 3 ohms. Increase power supply capacity or use SMPS. Voltage 18 volts. Output capacitor with low ESR.
Several smaller capacitors connected in parallel.
Voltage boost and feedback capacitors can not be changed (200 Hz).
Increase the driver current. Double the quiescent current compared to 3 ohms. Increase power supply capacity or use SMPS. Voltage 18 volts. Output capacitor with low ESR.
Several smaller capacitors connected in parallel.
Voltage boost and feedback capacitors can not be changed (200 Hz).
Last edited:
The problem with a low impedance speaker is that you need amps more than volts, but the increased current needs more drive power and/or output transistors.
the basic requirement is - what power do you need? then the volts=sqrt(PR) and amps=sqrt(P/R). 10W into 1 ohm therefore needs 3.16V and 3.16A rms.
You would best use transistors with high gain at high current, but quasi-saturation means having to use a higher voltage than ideal.
Note that in JLH's table, the distortion increases with reducing load. 1 ohm will be worse than the 3, assuming that the transistors you might use have gains of >100 as the ones JLH used (although modern devices should have better gain at higher currents, so will be rather better).
An output current of 4.5A peak (and that's only for 10W) will demand quite a high drive current, and may even need the driver to be doubled (Darlington or emitter follower boost), which will change the nature of the JLH original circuit.
But don't many ribbon speakers use matching transformers? You may be able to get Sowter to design/build an 8 -to-1 ohm high quality unit (but it won't be cheap.).
Possiblyy a JLH with MOSFEt outputs might be better placed for low impedances, but the low (depending on power needed) output voltage will still mean having to operate at a higher voltage than ideal - with a need for big heatsinks etc.
Main point - how much power do you need?
the basic requirement is - what power do you need? then the volts=sqrt(PR) and amps=sqrt(P/R). 10W into 1 ohm therefore needs 3.16V and 3.16A rms.
You would best use transistors with high gain at high current, but quasi-saturation means having to use a higher voltage than ideal.
Note that in JLH's table, the distortion increases with reducing load. 1 ohm will be worse than the 3, assuming that the transistors you might use have gains of >100 as the ones JLH used (although modern devices should have better gain at higher currents, so will be rather better).
An output current of 4.5A peak (and that's only for 10W) will demand quite a high drive current, and may even need the driver to be doubled (Darlington or emitter follower boost), which will change the nature of the JLH original circuit.
But don't many ribbon speakers use matching transformers? You may be able to get Sowter to design/build an 8 -to-1 ohm high quality unit (but it won't be cheap.).
Possiblyy a JLH with MOSFEt outputs might be better placed for low impedances, but the low (depending on power needed) output voltage will still mean having to operate at a higher voltage than ideal - with a need for big heatsinks etc.
Main point - how much power do you need?
Last edited:
Thick conductors, reinforcement of PCB tracks with copper core, good contacts.
I would think of a discrete Darlington at the output (with 2-4 power transistors at the output). The issue of thermal stabilization remains open.
Possible structure A + C. Above is the classic Darlington, and below is A + C. Resistor B-E of the lower output transistor 0.5-1 ohm.
Above may be a powerful MOS FET.
I would think of a discrete Darlington at the output (with 2-4 power transistors at the output). The issue of thermal stabilization remains open.
Possible structure A + C. Above is the classic Darlington, and below is A + C. Resistor B-E of the lower output transistor 0.5-1 ohm.
Above may be a powerful MOS FET.
Last edited:
Member
Joined 2009
Paid Member
Ah! this is my TGM8 amplifier front end more or less, I think this approach was discussed a few months back and I promised to start my own thread - well that hasn’t happened yet of course. It’s a great way to improve the performance. This is how I plan to rebuild my original version, although it will be all BJT.