You can add a zener diode to the circuit and get a stabilizer when the voltage is exceeded...
TO-3 transistors are supposedly labeled MJ2955 - You can install other suitable ones (p-n-p).
Last edited:
Regarding a stabiliser, when I mention stabiliser, I generally mean a feedback controlled power supply, not a capacitance multiplier. A capacitance multiplier acts as it says - more or less- as a larger capacitor than is used because of the gain of the transistor. The voltage on a cap. multiplier can still droop with current loading. It usually works well for a JLH though because the Class A circuit draws a constant current (on average, apart from when it might be overdriven). So you could use a multiplier in this case. Perhaps with an overvoltage Zener limiter as OldDIY suggests. (His illustration won't be enough current for your JLH though - you'll need a much beefier design, possibly with a transistor amplified zener too).
The MJ2955 is an epitaxial base PNP transistor, and although still fairly respectable it is of the older design and has gain droop- the data sheet shows a gain of 75 (typ.) at 1A but 50 at 2A and 30 at 4A. Let's say your power supply runs the amp at 20V. Keeping things simple the peak output voltage should be 10V, which needs 2.5A for 4 ohms or 3.3A for 3 ohms. You have to set the bias current to each transistor at half the base current needed for peak output (plus a little for a margin). Let's say that's a gain of 40 at 3A. The base current will be 75mA per transistor at peak output, so you have to set the bias to provide 37.5mA in each device in the quiescent state. The resistors in the bias chain have to conduct all of the base currents (75mA) which tells you what the total resistance (the two resistors in the collector of the VAS) needs to be. But at 37.5mA base current, the gain is going to be around the 50 mark, so the quiescent current won't be half the current you thought you needed from a basic calculation but nearer 2A, as OldDIY mentioned.
However, the main point is that your transistors might not have the typical datasheet value of gain. You need to be prepared to adjust the resistor values to get about a 2A quiescent current. The gain could be almost half, or double, pretty much.
Yes the input you asked about is lower for the 4 ohm version because you need a lower voltage for a given power into a lower resistance.
The downside is that while the input voltage is lower the current as has been discussed is higher - and that means the current swing in the circuit is higher, and that means higher distortion than the figures I think JLH implied.
The MJ2955 is an epitaxial base PNP transistor, and although still fairly respectable it is of the older design and has gain droop- the data sheet shows a gain of 75 (typ.) at 1A but 50 at 2A and 30 at 4A. Let's say your power supply runs the amp at 20V. Keeping things simple the peak output voltage should be 10V, which needs 2.5A for 4 ohms or 3.3A for 3 ohms. You have to set the bias current to each transistor at half the base current needed for peak output (plus a little for a margin). Let's say that's a gain of 40 at 3A. The base current will be 75mA per transistor at peak output, so you have to set the bias to provide 37.5mA in each device in the quiescent state. The resistors in the bias chain have to conduct all of the base currents (75mA) which tells you what the total resistance (the two resistors in the collector of the VAS) needs to be. But at 37.5mA base current, the gain is going to be around the 50 mark, so the quiescent current won't be half the current you thought you needed from a basic calculation but nearer 2A, as OldDIY mentioned.
However, the main point is that your transistors might not have the typical datasheet value of gain. You need to be prepared to adjust the resistor values to get about a 2A quiescent current. The gain could be almost half, or double, pretty much.
Yes the input you asked about is lower for the 4 ohm version because you need a lower voltage for a given power into a lower resistance.
The downside is that while the input voltage is lower the current as has been discussed is higher - and that means the current swing in the circuit is higher, and that means higher distortion than the figures I think JLH implied.
Thank you also for the thorough explanations to all questions @john_ellis.
Looking at the heatsinks and chassis of the following build, which I'm waiting for, do you (or anyone else) think from experience it can effectively handle 2A with a 22V PS? Biased at midpoint 11V that makes for 22Watts, whereas the build advertises "15W", so I'm a bit worried.
https://doukaudio.com/collections/a...power-amplifier-stereo-single-ended-audio-amp
I'm asking because I might still might be in a position to return it, naturally not after I've done the soldering rework of R1. But I could perfectly understand that nobody is in a position to judge this without trying, just asking for honest opinions.
Looking at the heatsinks and chassis of the following build, which I'm waiting for, do you (or anyone else) think from experience it can effectively handle 2A with a 22V PS? Biased at midpoint 11V that makes for 22Watts, whereas the build advertises "15W", so I'm a bit worried.
https://doukaudio.com/collections/a...power-amplifier-stereo-single-ended-audio-amp
I'm asking because I might still might be in a position to return it, naturally not after I've done the soldering rework of R1. But I could perfectly understand that nobody is in a position to judge this without trying, just asking for honest opinions.
Are you building a JLH PNP from a kit or buying a ready made amplifier which may or may not be a JLH?
You would do well to look up a book of basic electronics and amplifier operations. Things like power and RMS seem to have escaped you, which is what you need to answer your questions.
As regards the image you showed, the heatsinks look like 200mm x 25 or 40mm deep, only 50 or 75 mm long. Maybe it has a thermal resistance in the region of 1C/W. Could be lower, if 40mm deep fins. After you've dug up some info about power and amplifier outputs, tell us what you think the total power input per channel is going to be and the depth of the heatsink fins. You can look up manufacturer's data as well as I to find out its thermal resistance and then work out how hot it will get.
You would do well to look up a book of basic electronics and amplifier operations. Things like power and RMS seem to have escaped you, which is what you need to answer your questions.
As regards the image you showed, the heatsinks look like 200mm x 25 or 40mm deep, only 50 or 75 mm long. Maybe it has a thermal resistance in the region of 1C/W. Could be lower, if 40mm deep fins. After you've dug up some info about power and amplifier outputs, tell us what you think the total power input per channel is going to be and the depth of the heatsink fins. You can look up manufacturer's data as well as I to find out its thermal resistance and then work out how hot it will get.
You can find the leading dimensions for the case/heatsinks here: https://www.aliexpress.com/item/328...7!sea!AU!178186480&curPageLogUid=0UVZwv9Ymsyh
Thanks. According to a supplier that's 1.65C/W if fins are 25mm deep but in black, anodised.
I'm more used to using older TO3 types which were usually about 4 and wondered what to use. Not done the sums but these sound feasible.
Hello Chumingo,
My speakers are Mark Audio CHP-70's 4 Ohm 85 db
I use a 12V toroid and get roughly 5 Watts max out of it.
What I actually listen to at a loud volume is 1/32nd of a Watt with a worst case 18 db transient I might need 2 Watts
My heatsinks get warm not hot.
The sound is dramatically; fast, bright, detailed, analytical and transparent. And this all comes from a £10 Chinese kit.
It's an electronical joke
That so much can come from something so small.https://jlh1969classa.blogspot.com/p/my-build.html
Cheers - J
Last edited:
Hello Chumingo,
My speakers are Mark Audio CHP-70's 4 Ohm 85 db
I use a 12V toroid and get roughly 5 Watts max out of it.
What I actually listen to at a loud volume is 1/32nd of a Watt with a worst case 18 db transient I might need 2 Watts
My heatsinks get warm not hot.
The sound is dramatically; fast, bright, detailed, analytical and transparent. And this all comes from a £10 Chinese kit.
It's an electronical joke
https://jlh1969classa.blogspot.com/p/my-build.html
Cheers - J
Thanks for the info @jemraid, actually I'm not seeking for loud volume this is mostly to monitor environmental audio tracks, so it might work.
Today the amp arrives
And 4ohm speakers are waiting for it. Only downside is the included 22V toroid, so I might have to swap it for a 17V or 12V as you did... still studying and thinking about it.@john_ellis It took me a while to realize watts were being measured in RMS.
Happy to hear about the 22th anniversary!!! Holy c***
The power goes to 2 transistors. For 10w it seems the power dissipated by each transistor will be ~17w based on Hood's articles. That tends to apply what ever the load is as the supply voltage is changed for each of the speaker impedances. Eg he suggests 17v 2amp for 3ohm, 2amp and 27v for 8,1.2amp. 3ohm works out at 34w, 8ohm, 32w. Rather similar.
For your 15w you can simply scale the numbers accounting for a 4ohm load and the increased power. It will be close enough for your purposes. Powere disipation in each transistor should be ~17x1.5 watts. Neither transistor's current can go down to zero so there has to be an excess.
I bought the white boards with the T0-3-0 outputs (were pulls). One tested bad and lucky enough & I found a good one in my parts draw. These are mounted on a angle bracket with tiny odd insulators. Since I'm running a low voltage, I don't see a problem, but rather funky. I'm waiting for a transformer to arrive and make some kind of chassis. Always surprises with this stuff. But I thank the forum your sharing their info.
Markings are 100% and £2.50 a piece from an approved dealer.
interesting choice of part.
It's a bit late to add this, but surely most of you will have compared what you can see in the above pics of the CIRMECH board and its little heatsink, with other typical JLH class A kits, finished products and boards. That heatsink is nowhere near large enough for more than about 4W per channel in class A. It is woefully undersize and you'll wind up needing to pack it in ice before it overheats and your semis burn, genuine or not, if your power supply is more than about 15V.
I have one of these too, and realised how far short of a JLH amp it was before trying to power it. I've estimated that it needs a heatsink at least 4x that size to peak at 15W/channel before clipping and overheating when the ambient temp. is 20-25C. Perhaps bolting the 'sink to a heavy aluminium chassis or a much larger heatsink would allow it to deliver what you could have expected.
The TTC5200 transistors appear to be genuine Chinese assembled types but TIP41 phase splitter/driver transistors are not a good choice at all (high capacitance, very low Ft and hFE). If you obtain an original TO5 type like 2N1711, so much the better but 2SD669A (most Chinese versions are fine) is cheap, available, apparently better and already used on other commercial JLH clones. It should just drop in if the leads are splayed and it's rotated 180 deg. with respect to the TIP41's heatsink side. i.e. the pin order for TO126 style transistors is the reverse of TO220.
I have one of these too, and realised how far short of a JLH amp it was before trying to power it. I've estimated that it needs a heatsink at least 4x that size to peak at 15W/channel before clipping and overheating when the ambient temp. is 20-25C. Perhaps bolting the 'sink to a heavy aluminium chassis or a much larger heatsink would allow it to deliver what you could have expected.
The TTC5200 transistors appear to be genuine Chinese assembled types but TIP41 phase splitter/driver transistors are not a good choice at all (high capacitance, very low Ft and hFE). If you obtain an original TO5 type like 2N1711, so much the better but 2SD669A (most Chinese versions are fine) is cheap, available, apparently better and already used on other commercial JLH clones. It should just drop in if the leads are splayed and it's rotated 180 deg. with respect to the TIP41's heatsink side. i.e. the pin order for TO126 style transistors is the reverse of TO220.
Last edited: