Ever heard of the skin effect on wire with frequency?? 🤣
Copper is better. The resistivity* of gold is 2.214E-8 ohm meters at room temperature. The resistivity of copper is 1.72E-8 ohm meters. So a gold wire has about 29% higher resistance than the same size copper wire.
What is really needed is lots of fine strands and even that has issues in the same area,
Last edited:
Not wipe away his listening experience (with all the context unknown to us) with a simple "theory";-)
So far, I prefer, at moment, single-core stripping. I also have an extremely good multi-core (each with insulated cables) to pass around. By far the best I have heard so far. But it "scales" the sound minimally. It is probably the minimally different lengths of the insulated wires that lead to this unrest in the sound. That's why I prefer to use my single-core cables, even if they don't sound nearly as lively, black, deep and contoured.
So far, I prefer, at moment, single-core stripping. I also have an extremely good multi-core (each with insulated cables) to pass around. By far the best I have heard so far. But it "scales" the sound minimally. It is probably the minimally different lengths of the insulated wires that lead to this unrest in the sound. That's why I prefer to use my single-core cables, even if they don't sound nearly as lively, black, deep and contoured.
More topical. I came across an amp design that had a particular style of filtering for stability. I've seen it on one other as well but on this one the filter was before the feedback network at the output making the reason more obvious. 2w 10R resistor and some nF down to ground. Bandwidth was still 350kHz,
This design was using an OP amp. The network IMHO is to take out frequencies where phase shift in the output stages would cause problems.Maybe something to try if a Hood misbehaves. It would be hung on the output.
This design was using an OP amp. The network IMHO is to take out frequencies where phase shift in the output stages would cause problems.Maybe something to try if a Hood misbehaves. It would be hung on the output.
Cable sellers would have you believe otherwise but the Skin Effect has little influence at audio frequencies, depending to some degree on the cable length and signal impedance. In other words, keep your cables as short as possible, of lowest practical resistance and you will never need to consider it. The only downsides of cheap cables and signal leads is their appearance, toughness and maybe they just don't have a suitable specification for use with all types of home audio.
The tiny cables used for cheap headphones, buds and interconnect cables are suspect but short, fat speaker cables or a 30cm long interconnect lead? Not in my experience and apparently, many others too. e.g. https://www.soundguys.com/best-audiophile-cables-21871/
The tiny cables used for cheap headphones, buds and interconnect cables are suspect but short, fat speaker cables or a 30cm long interconnect lead? Not in my experience and apparently, many others too. e.g. https://www.soundguys.com/best-audiophile-cables-21871/
Commonly known as a Zobel network. Most medium to large power amps with substantial gain and feedback, should have one. However, articles on the web are at variance as to what their main function is. Snubber? HF damper or load? Neutralize voicecoil inductance effects? All of these? Well, they are intended to counter the rising load impedance at higher frequencies to ensure stable operation across the audio band under varying operating conditions and that's worth the cost of repairing amplifiers, any time.
Last edited:
![DSCN0289[1].JPG](/community/data/attachments/1081/1081240-bead9caff30f074653bc7fec5d5b5f4f.jpg?hash=vq2cr_MPB0)
Not entirely true. Much noise was made about it when initial switch mode power supplies were at 20khz 😉 so people couldn't hear them. Many could, A decent audio amp is likely to prduce half decent square waves at that frequency but the effect on loudspeaker leads wont be worth worrying about.
As said, "little influence". It can't be concluded though, that the many people who heard the effects of a 20kHz SMPS could attribute it to skin effect or any particular mechanism. It only says that many people could detect spurious sound (likely as a sub-harmonic or product of the 20 kHz switching frequency and/or associated harmonics and noise) of the SMPS.
A spectrum analyser would have been appropriate and indicated something closer to the truth but the solution is likely in redesigning the amplifier such that it has better suppression of EMI and noise in its filtering/shielding. Sure, any audio PSU switching frequency should be well above 20kHz for use near audio equipment and preferably have power factor correction (PFC) too, as generally found in better SMPS units. This also contributes to noise suppression.
A spectrum analyser would have been appropriate and indicated something closer to the truth but the solution is likely in redesigning the amplifier such that it has better suppression of EMI and noise in its filtering/shielding. Sure, any audio PSU switching frequency should be well above 20kHz for use near audio equipment and preferably have power factor correction (PFC) too, as generally found in better SMPS units. This also contributes to noise suppression.
Last edited:
Copper corrodes. Some corrosion products are semiconducting. Ask any (sweaty) guitar player who's near an AM radio station and used a crimped (not soldered) connector. Litz wire and careful soldering can fix most of those issues.
There's all sorts of compromises in cable & connector design & manufacture: plating brings it's own set of charms. Silver is lovely, if money is no object.
Some times facts related to physics and other factors creep into the wider world and get seriously distorted. One aspect mentioned is associated with dissimilar metals. The other relates to human hearing.
I am not a moderator but allow me to tell you that this thread has seriously drifted, it would be good for everyone if we returned to the original subject so as not to further pollute this already well-polluted thread.
Hello there,
I concur HG
I've got a cheapo Chinese kit that sounds great but as one does I found some odds and ends. Starting with the Blue RS steel box from a car boot sale, a toroid and a some caps.
I made another front plate and put the bits on a plywood base so I could slide in into the box.
Another find was 4 MJ15003's in my rubbish box and I've put in a bid for a couple of heatsinks similar to JLH's.
More eventually
Cheers - J
I concur HG
I've got a cheapo Chinese kit that sounds great but as one does I found some odds and ends. Starting with the Blue RS steel box from a car boot sale, a toroid and a some caps.
I made another front plate and put the bits on a plywood base so I could slide in into the box.
Another find was 4 MJ15003's in my rubbish box and I've put in a bid for a couple of heatsinks similar to JLH's.
More eventually
Cheers - J
That looks like a nice project in the making. Everything looks just fine for the period but the heatsinks you plan on, are designed to mount vertically (i.e. with the fins aligned vertically) so that the hot air flows freely and escapes upwards by convection. We're also looking downward on JLH's project but his case is at least some 5" (~ 13cm)` high and shown open at the top, which would approach optimum for air flow, as mentioned earlier.
You will certainly need a lot more cooling capacity than any 10W AB class amp - think of something equivalent in heat output to a typical 60-80W/channel stereo amplifier when it's working hard. I suspect that if you mount the original type of heatsinks sideways to fit them in, the air flow will be seriously impaired due to poor flow to and from the fins and without large openings in the top and bottom of the case.
You will certainly need a lot more cooling capacity than any 10W AB class amp - think of something equivalent in heat output to a typical 60-80W/channel stereo amplifier when it's working hard. I suspect that if you mount the original type of heatsinks sideways to fit them in, the air flow will be seriously impaired due to poor flow to and from the fins and without large openings in the top and bottom of the case.
Last edited:
As Ian mentioned you are going to have to consider the heatsink carefully. You can buy low Rth heatsinks with a low profile but they will have deep fins, a sort of one-sided version of the Marston 10D type. You might need to cut big windows in the top and bottom for good heat flow, if the heatsinks are inside the box, or, depending on how the rest of the circuitry fits in, you may need to hang them out of the back. They should be mounted fins vertically, as Ian mentioned, as they will get hot. Dissipation of a 10W JLH is 17W/transistor or 34W per channel (nominally, for 27V/1.2A - but I have suggested (several times) that you may need to run at a higher voltage for lower distortion near full power output due to quasi-saturation). On the old 10D heatsink that was 2.2C/W for a 4 inch length or 1.5C/W for a 6 inch length (single devices centrally mounted so in practice a little better with two devices) (IIRC - I don't have access to my data files at the mo.) - the 4 inch /100mm would see a 75C rise above amb. which is above the temp recommended these days without secondary protection (i.e. enclosure) to prevent finger burning (and an enclosure probably means higher temperatures!) You would really need a 2.2C/W per transistor or < 1C/W for both.
An enclosure could of course be a wire type cage, or similar grid, a suitable distance outside the heatsinks, which would not impact air flow significantly.
An enclosure could of course be a wire type cage, or similar grid, a suitable distance outside the heatsinks, which would not impact air flow significantly.
it would be good for everyone if we returned to the original subject so as not to further pollute this already well-polluted thread.
FWIW, In 1970, Nelson Jones followed up on JLH's amplifier with his own, more conventional class A design: https://sound-au.com/tcaas/nj10w.pdf. https://www.diyaudio.com/community/threads/nelson-jones-10w.290890/
I have no idea of the audio quality or popularity of his DIY design but there is a discussion and details of the heatsinks in the article (4 required for stereo) that have a similar duty to JLH's there, which may be helpful.
I have no idea of the audio quality or popularity of his DIY design but there is a discussion and details of the heatsinks in the article (4 required for stereo) that have a similar duty to JLH's there, which may be helpful.
Last edited:
He added a tracking bias node to the class AB quasi-complementary amplifier. I have never tested suitability. The output of the BJT worked in class A. The calculations of the modes for various load resistances, which may be of interest, are given.
Nelson Jones amp didn't cause much of a reaction in the UK at the time. More complication compared with JLH's may be the reason and in some respects not much gain.
Hetsinks now are rather different to what was available then. 2 options available in the UK. The search isn't very good on this site and TO3 may not show this one if used in a search
https://cpc.farnell.com/abl-heatsinks/515ab0750mb-t03/heat-sink-to-3-1-9-c-w/dp/SC10874?st=heatsink
😉 TO3 just means drill one or more holes to hold the transistors. It doesn't seem to be predrilled. There will be other similar ones on the site
Another type is this
https://uk.rs-online.com/web/p/heatsinks/9033087
More trouble to use as the drilled hole needs to be tapped with simple tools to provide a screw thread and only suitable for plastic power transistors. The hole also needs to be pretty square to the surface but a plain and spring washer could help with that but the threads will accommodate a bit of tilt, probably enough with a bit of care.
To suite a case the length needs to be suitable to finish up with vertical fins. Farnell are the parent company of CPC and their search may be better, length range selected in searches as per RSWW. Last time I ordered off them it was free postage but they added a £6 handling fee. RSWW - just charge postage. Too many small customers to continue with free postage. CPC can be better for toriods as all include the mounting kit.and provide data,
Hetsinks now are rather different to what was available then. 2 options available in the UK. The search isn't very good on this site and TO3 may not show this one if used in a search
https://cpc.farnell.com/abl-heatsinks/515ab0750mb-t03/heat-sink-to-3-1-9-c-w/dp/SC10874?st=heatsink
😉 TO3 just means drill one or more holes to hold the transistors. It doesn't seem to be predrilled. There will be other similar ones on the site
Another type is this
https://uk.rs-online.com/web/p/heatsinks/9033087
More trouble to use as the drilled hole needs to be tapped with simple tools to provide a screw thread and only suitable for plastic power transistors. The hole also needs to be pretty square to the surface but a plain and spring washer could help with that but the threads will accommodate a bit of tilt, probably enough with a bit of care.
To suite a case the length needs to be suitable to finish up with vertical fins. Farnell are the parent company of CPC and their search may be better, length range selected in searches as per RSWW. Last time I ordered off them it was free postage but they added a £6 handling fee. RSWW - just charge postage. Too many small customers to continue with free postage. CPC can be better for toriods as all include the mounting kit.and provide data,
to mount TO-3's on a flat based heatsink that can be done using a length of e.g. 35mm aluminium angle 5 or 6mm thick. Mount the To-3's on one side and bolt it to the heatsink with e.g. M5 screws with the other. Better results than tapping M3 which plastic devices generally require. Takes more space - but that might not be too much of an issue as it will permit better air flow.
Summarising Nelson-Jones's comments on heatsinks, he used a 4 inch (100mm) length of what looks like Marston 10D (hard to find these days but there are similar ABL), 2.2C/W with one per transistor directly mounted (no insulator).
Depending on the size of the OP's box heatsinks may need to be extending out of the rear or even on top.
You can mount heatsinks which have to be insulated using a small aluminium angle and additional M3 (though M3.5 would be sturdier) insulating washers and M3 (M3.5) screws if the heatsinks have flanged sides. If not, you would need to use long M3 (M3.5) screws or cut lengths of threaded rods to mount the heatsinks using the main centre section with insulators and nuts/washers each side.
There are single-sided (flat based) heatsinks similar to 10D (around 2C/W in the same size (100mm wide and long)) which have a missing fin or two in the centre to allow mounting TO-3's from the finned side (drill-your-own).
Worst case, use plain insulating washers (e.g. fibre) under metal washers and insulate the rods or screws with sleeving inside the holes they pass through. If you mount 5(???) 10D types in a stack with long threaded rods you may need M4 size for strength, passing through fins to pick up the centre section (difficult to assemble - and to be able to tighten nuts only m3.5 might be feasible to get spanners in between fins)
Summarising Nelson-Jones's comments on heatsinks, he used a 4 inch (100mm) length of what looks like Marston 10D (hard to find these days but there are similar ABL), 2.2C/W with one per transistor directly mounted (no insulator).
Depending on the size of the OP's box heatsinks may need to be extending out of the rear or even on top.
You can mount heatsinks which have to be insulated using a small aluminium angle and additional M3 (though M3.5 would be sturdier) insulating washers and M3 (M3.5) screws if the heatsinks have flanged sides. If not, you would need to use long M3 (M3.5) screws or cut lengths of threaded rods to mount the heatsinks using the main centre section with insulators and nuts/washers each side.
There are single-sided (flat based) heatsinks similar to 10D (around 2C/W in the same size (100mm wide and long)) which have a missing fin or two in the centre to allow mounting TO-3's from the finned side (drill-your-own).
Worst case, use plain insulating washers (e.g. fibre) under metal washers and insulate the rods or screws with sleeving inside the holes they pass through. If you mount 5(???) 10D types in a stack with long threaded rods you may need M4 size for strength, passing through fins to pick up the centre section (difficult to assemble - and to be able to tighten nuts only m3.5 might be feasible to get spanners in between fins)
Last edited: