Heat sink drilling for TO3 is fun.
I use the pcb as a template taking care not to enlarge the holes.
I drill the hole at one end of the heatsink then bolt it to the pcb.
I then drill the hole at the opposite ned of the heatsink and bolt that to the pcb.
You can then drill the other holes.
Make sure the holes are deburred or the burr will break through isolation tabs.
I use the pcb as a template taking care not to enlarge the holes.
I drill the hole at one end of the heatsink then bolt it to the pcb.
I then drill the hole at the opposite ned of the heatsink and bolt that to the pcb.
You can then drill the other holes.
Make sure the holes are deburred or the burr will break through isolation tabs.
An externally hosted image should be here but it was not working when we last tested it.
Yes its just bog standard ally angle.
I got mine off ebay, they even cut it to length for you if you ask.
Thanks Nigel, that's easy then. I'm shopping for caps, I need a pair of 10,000 35V (32V trafo) for the PSU, a pile of 220uF and a pair of 100uF. May as well go 35V on all. The catalogues talk about "snap in", what does that mean? Is it for SMD boards? Will they work on a standard through-board PCB like mine?
Edit - how do you isolate the underside of the ally bracket from the PCB? Anything clever, or just a plastic membrane?
Edit - how do you isolate the underside of the ally bracket from the PCB? Anything clever, or just a plastic membrane?
Last edited:
I made the holes not through hole plated so there is no copper on the top of the pcb, just the bottom
If you have copper on top of the pcb you will need to add a plastic membrane.
Snap in are just through hole capacitors.
https://uk.rs-online.com/web/p/aluminium-capacitors/8712717/
If you have copper on top of the pcb you will need to add a plastic membrane.
Snap in are just through hole capacitors.
https://uk.rs-online.com/web/p/aluminium-capacitors/8712717/
The letter code is a European standard identifier which tells us the type and construction of film capacitors. Nowadays, many new film types and forms of construction with different, sometimes critical virtues, have become common in supplier catalogues. Even DIY specifications will need to include this for major distributors who carry a wide range covering many types and formats like SMD, not just the cheapest available through-hole types of many years ago when we bought from smaller stockists and retail outlets. I was just as unaware of new developments then, as DIY projects always leaned toward the common and cheapest available parts.
Here, M shows it's a vapour deposited, metallized film type, K that it is in a standard, rolled form of 2 offset film layers, and P that the film material is polypropylene. Common polyester caps are identified as MKT (T for poyethylene teraphthalate or PET plastic, as we now call it). Anyway, most of what you may need to know about different film capacitor types and reading values is here: Capacitor code table | Kaizer Power Electronics
BTW, though its what you commonly find, 3mm angle aluminium won't be good enough to conduct the heat efficiently from even a 10-15 W class A amp's output transistors, where maximum heat dissipation is always on and you need large, solid heatsinks. The temperature differential between transistor case and and the heatsink will be a worry. Even the cheap Chinese kits supply 4-5mm extruded angle as necessary, which tells you something, as does the thickness of the heatsinks you'll need.
"Snap-ins" are intended to snap into 2.0 dia. holes spaced at 10 mm centres and hold the cap on the PCB during soldering. You could treat them like wires and straighten the kinks but I can't see why there would be any need (unless its a bodge) if the PCB is designed for modern components.
Go to the index of large can types and select the snap-in terminal types, terminal shape then PDF 1/2 here: NICHICON CORPORATION | Product & Technology
Edit: I'm not sure that boards for JLH'69 will suit Snap-ins, as the output and input caps are usually an axial leaded type, to keep the authentic appearance, I imagine. The power supply or perhaps a capacitance multiplier/filter (if included) could be a different matter. If you bought PCBs recently, how about a link to the product on their website or Ebay etc?
Go to the index of large can types and select the snap-in terminal types, terminal shape then PDF 1/2 here: NICHICON CORPORATION | Product & Technology
Edit: I'm not sure that boards for JLH'69 will suit Snap-ins, as the output and input caps are usually an axial leaded type, to keep the authentic appearance, I imagine. The power supply or perhaps a capacitance multiplier/filter (if included) could be a different matter. If you bought PCBs recently, how about a link to the product on their website or Ebay etc?
Last edited:
As said the Panasonic FC range of capacitors are well liked. To make clear, The output cap is a simple elegant solution to a problem. It makes the amplifier super safe. I have real doubts that anyone hears any real distortion caused by them. I consider myself to be reasonably good at subjective evaluation, doing tests with direct coupled verses capacitor coupled led me to think it was suprisingly hard to tell which was which. The test used 4000 uF of Nikkia non polar caps on an amplifer with +/- power rails very like a JLH.
No expert on this forum learnt without making mistakes. If some caution is exercised the input capacitor might be removed. The preamp must have a capacitor output as must a volume contol. If in doubt measure the DC voltage at the amplifier output cap before connecting the preamp, allow 5 minutes warm up period. If it changes by more than 0.5V when preamp connected it would be wise to fit an input cap. I always fit one as I never know what I might do next. I suspect I would fit 2u2 250V polyester.
You might be surprised to know that the volume control can change the sound more than often suggested improvements. My best guess is 10K log type will be OK. The more common 20K should be fine. 5K also. This assumes the preamp can drive that control. A CD player should drive it. At a pinch a CD should go loud enough. If it doesn't the gain of the amplifier can be changed. As it is class A and remakably low distortion for it's type changing the gain resistor to 100R from 220R will more than double the gain. Distortion will not be greater than the JLH design criteria of 0.1% THD. I and others got 0.05% when standard. I suspect it is lower if I had the tools to say.
This part of hi fi theory isn't talked about much. That is matching the volume control to the amplifier input. To repeat myself. If only a volume control is used ( often called a passive premap ) it might cause the amplifier to sound weak or muffled. Equally it can sound better than any active preamp. If the control is too high in value the current to drive the JLH input transistor might be a bit low. I used BC327-40 to ensure that would be unlikely, although it is the 100K in paralell with circa 120K that is the main question. If the volume control is very low resistance this could be ideal ( remember the amplifier must have an input cap for this idea ). The results should be very open sound with very little loss of detail due to hiss. The problem might be if it can be driven. What I do is make all my driving devices have enough current. The ability to drive 600 ohms is not a bad design principle which is often stated . Even some 3 transistor types will to 1 Vrms ( ESP Audio often are a source or TubeCad who isn't just tubes ). One seldom stated fact is a " passive preamp " in theory has infinite overload margine. That must be the most important hi fi criteria never talked about even if not ideal for a guitar amp.
Often active 318/3180 uS + 75 uS passive is a type of phone stage that can work best. It will have about 30 dB overload margine which should be fine ( 10 dB if cartridge suited to the gain works, science says the same ). If the 318/3180 is the output stage and has a gain of about 16 at 1kHz ( circa 160 at bass end ) there sould be enough current to drive the vollume control. A NE5532 should be OK and a MC33078 better. The input op amp doing the 75uS and a gain of 4 might suit. A Tuner might need a gain of 3.
No expert on this forum learnt without making mistakes. If some caution is exercised the input capacitor might be removed. The preamp must have a capacitor output as must a volume contol. If in doubt measure the DC voltage at the amplifier output cap before connecting the preamp, allow 5 minutes warm up period. If it changes by more than 0.5V when preamp connected it would be wise to fit an input cap. I always fit one as I never know what I might do next. I suspect I would fit 2u2 250V polyester.
You might be surprised to know that the volume control can change the sound more than often suggested improvements. My best guess is 10K log type will be OK. The more common 20K should be fine. 5K also. This assumes the preamp can drive that control. A CD player should drive it. At a pinch a CD should go loud enough. If it doesn't the gain of the amplifier can be changed. As it is class A and remakably low distortion for it's type changing the gain resistor to 100R from 220R will more than double the gain. Distortion will not be greater than the JLH design criteria of 0.1% THD. I and others got 0.05% when standard. I suspect it is lower if I had the tools to say.
This part of hi fi theory isn't talked about much. That is matching the volume control to the amplifier input. To repeat myself. If only a volume control is used ( often called a passive premap ) it might cause the amplifier to sound weak or muffled. Equally it can sound better than any active preamp. If the control is too high in value the current to drive the JLH input transistor might be a bit low. I used BC327-40 to ensure that would be unlikely, although it is the 100K in paralell with circa 120K that is the main question. If the volume control is very low resistance this could be ideal ( remember the amplifier must have an input cap for this idea ). The results should be very open sound with very little loss of detail due to hiss. The problem might be if it can be driven. What I do is make all my driving devices have enough current. The ability to drive 600 ohms is not a bad design principle which is often stated . Even some 3 transistor types will to 1 Vrms ( ESP Audio often are a source or TubeCad who isn't just tubes ). One seldom stated fact is a " passive preamp " in theory has infinite overload margine. That must be the most important hi fi criteria never talked about even if not ideal for a guitar amp.
Often active 318/3180 uS + 75 uS passive is a type of phone stage that can work best. It will have about 30 dB overload margine which should be fine ( 10 dB if cartridge suited to the gain works, science says the same ). If the 318/3180 is the output stage and has a gain of about 16 at 1kHz ( circa 160 at bass end ) there sould be enough current to drive the vollume control. A NE5532 should be OK and a MC33078 better. The input op amp doing the 75uS and a gain of 4 might suit. A Tuner might need a gain of 3.
Hi Steve,
How is your '32v trafo' configured ? 16VAC-0-16VAC? Then 35 v caps are ok for a dual supply giving approx +/-22VDC
Brilliant, thanks gents, there is loads to be going on with there. I will add the schematic/link to PCB later. It's copper both sides and through-hole plated so I will need a gasket below and sleeves on the E and B pins where they pass through. It's occurred to me overnight that a paper/card gasket will suffice to isolate the bracket from the board underside, as would a plastic sheet. It also has the Darlington pair capacitance multiplier built in.
Point noted re ally thickness, I will have a look in the workshop, I may be lucky. I do appreciate that the difficulty will be in getting the heat from the components to the bracket to the heat sink, a 10°C differential at each step means that a 50°C fin could be a 60°C heat sink body, a 70°C bracket, an 80°C TO3 case, and Gawd knows what at the junction that needs to be cooled.
I have a further noob question, re trannies. The schematic calls for one PNP device, a 2N2907. It's in a TO-18 case as you might imagine for a conventional device in 1969. hFE is 100-300. I have had a rummage in the bits box and turned up some unidentified 1980's PNP signal trannies in plastic cases, no codes. Measurement shows a pair that show hFE 150, so middle of spec, would it be stupid to give these a try? I hear the JLH isn't choosy about semiconductor type, and the duty doesn't look severe, is it worth a shot or a case of "Look soft lad, don't run before you walk, buy the right item then there is a fighting chance that it might work"?
@Prasi, yes, it's a 16-0-16 from an old AV amp that a pal donated for the purpose. Quoted input power 140W so big enough I think. It gives me case and trafo for free, can't be bad. Free heat sink too from another friend, it's cost pennies to date!
Point noted re ally thickness, I will have a look in the workshop, I may be lucky. I do appreciate that the difficulty will be in getting the heat from the components to the bracket to the heat sink, a 10°C differential at each step means that a 50°C fin could be a 60°C heat sink body, a 70°C bracket, an 80°C TO3 case, and Gawd knows what at the junction that needs to be cooled.
I have a further noob question, re trannies. The schematic calls for one PNP device, a 2N2907. It's in a TO-18 case as you might imagine for a conventional device in 1969. hFE is 100-300. I have had a rummage in the bits box and turned up some unidentified 1980's PNP signal trannies in plastic cases, no codes. Measurement shows a pair that show hFE 150, so middle of spec, would it be stupid to give these a try? I hear the JLH isn't choosy about semiconductor type, and the duty doesn't look severe, is it worth a shot or a case of "Look soft lad, don't run before you walk, buy the right item then there is a fighting chance that it might work"?
@Prasi, yes, it's a 16-0-16 from an old AV amp that a pal donated for the purpose. Quoted input power 140W so big enough I think. It gives me case and trafo for free, can't be bad. Free heat sink too from another friend, it's cost pennies to date!
Last edited:
One thing you could try is leaving space under the PCB. Try double capacitors. That is if 220uF make it 440 uF by soldering to the PCB back ( + to +, - to - ). See which sound you like best. Less can be more. It can give tighter bass. Usually it goes the other way. If worried about sub bass and LP warps the output caps can be standard JLH, assume that for now. However sub bass is sometimes exactly what you want ( 5 Hz ). If so have a LF filter option on phono. Even then sometimes cones flapping still sounds best. Sub bass is the ghostly quality and is nothing to do with banging bass. It gives mid range tone colour and is well known to organists. OK the notes stop short of that, it still makes a difference. A Jesuit priest taught me that and used real music to show it. Real ghostly feeling are most likely due to that. It is said Wind Farms can make people think there are ghosts. We can't hear it yet we know it's there. JLH liked Quad ESL 57's. Although they go to 40Hz they can show sub bass, it's the note shape.
For the PNP I would buy new as state of the art can be 6 pence.
For the PNP I would buy new as state of the art can be 6 pence.
I hoped to do this as an edit but it won't let me just now, so sorry for the additional post.
The PCB has drillings of 1mm at 6mm centres for the 220uF caps, and at 4mm centres for the 100uF. The 10,000uF on the PSU has multiple drillings, 2mm, at 10, 17, 23 mm centres, so that should be easy, it also needs an output cap, 2200uF, that has 1mm holes at 6,7,8 mm or an option to have the can horizontal with a connection at each end, seems easy enough.
Edit, thanks Nigel, I shall have an outbreak of common dog and buy the actual trannies, 2 off 2N2907. Then we know. 5 are £1.24 at Farnell, not the price of the shoe leather to get there.
The PCB has drillings of 1mm at 6mm centres for the 220uF caps, and at 4mm centres for the 100uF. The 10,000uF on the PSU has multiple drillings, 2mm, at 10, 17, 23 mm centres, so that should be easy, it also needs an output cap, 2200uF, that has 1mm holes at 6,7,8 mm or an option to have the can horizontal with a connection at each end, seems easy enough.
Edit, thanks Nigel, I shall have an outbreak of common dog and buy the actual trannies, 2 off 2N2907. Then we know. 5 are £1.24 at Farnell, not the price of the shoe leather to get there.
Last edited:
I have a question too - about the power supply. I doubt it's Hood's design as whilst 2N2907 /2N2222 GP types were very common and a dime a dozen in the US, they weren't seen much in UK designs that I recall. So where is the design? Just post a link to it so we can see what it is you want to do. You can usually substitute those metal can types with their plastic case versions, PN2907 / PN2222. Also if still around, PN200/PN100 would likely work fine too and either way, if they come from reputable suppliers, you needn't worry about hFE.
What I do is use various caps to my speaker units. That is 4700 uF to the 15 inch bass and 1000 uF to the 12 in fullrange with about 2u4 to the tweeter. The bass is 5.7 ohms so 6 Hz and the 12 inch 8 ohms and 20 Hz. The tweeter 8 ohms 8300 Hz. This remove a big critisisum against the use of an output cap. The tweeter is no different to any other speaker so can't be worse and is actually better. The 1000 uF allow only the listener to be damaged as 115 dB is easy to have. It is also vital to do as the speaker is on a baffle, it gives a good compromise. The bass unit will take 500 watts and needs 10 watts. It has a 7.5 mH choke to roll the mid range off that starts about 120 Hz ( it must be lower inductance as it higher than that ). When time allows I will get a full JLH working to save mine and Colleen's ears. Even then 108 db will be easy to get.
I bought this PCB:
2pcs JLH 1969 class A amplifier amp stereo high quality bare PCB 10W DIY audio | eBay.
The schematic is a PDF below.
In the last photo in the Ebay pic there is a populated board, it has green connectors for Line in, load out and power. What are these called in electronics catalogues?
Last edited:
I suggest you will need at least 3mm ally for the bracket. Taking an example bracket 50mm per side, 150mm long (yours might be 200?) then a quick calculation of the thermal resistance for a transistor mounted in the middle (25mm centre) to the centre of the other arm works out to be about 1.2K/W, though heat will start transferring of course once past the right angle bend, so should be better. Assuming a dissipation of 15W per transistor that is still quite a temperature rise, but that is the Class A trade-off!
I've used 3mm for a 50W amp (class B) without issues, with the transistors dissipating about the same at near full power.
I've used 3mm for a 50W amp (class B) without issues, with the transistors dissipating about the same at near full power.
I think I would buy these kit if ever would build these amp.
PNP sanken A1216 JLH1969 single-ended class A power amp kit 10W+10W L168-27 | eBay
PNP sanken A1216 JLH1969 single-ended class A power amp kit 10W+10W L168-27 | eBay
Thanks for the schematic and pic links. That looks like some high quality boards shown here recently. The power supply is a straightforward bridge rectifier and electrolytic cap followed by a capacitance multiplier circuit which further filters the DC with less voltage drop than a regulator. 'Good idea and one JLH suggested in the original article.
It spoils the shopping adventure, but you can buy the kits with all quite good components or finished and working too, for less than you'll spend on similar quality parts at local suppliers. Semiconductors can be a problem with sometimes poor copies, substitutes and occasional fakes from China but typically, their kit and assembled quality is fine and unsurpassed value.
The board is intended for a "snap in" 10,000 uF electrolytic and as mentioned, the output cap. is an axial type (the leads are arranged on its axis) and the input cap footprint would only suit a radial lead type (the leads are one the same face). However, all I've seen are axial MKP type with leads bent to suit.
It spoils the shopping adventure, but you can buy the kits with all quite good components or finished and working too, for less than you'll spend on similar quality parts at local suppliers. Semiconductors can be a problem with sometimes poor copies, substitutes and occasional fakes from China but typically, their kit and assembled quality is fine and unsurpassed value.
The board is intended for a "snap in" 10,000 uF electrolytic and as mentioned, the output cap. is an axial type (the leads are arranged on its axis) and the input cap footprint would only suit a radial lead type (the leads are one the same face). However, all I've seen are axial MKP type with leads bent to suit.
The boards are of very high quality, they are lovely. I have found a pair of snap-in 10,000uF caps in the bits box, I think they were from an amplifier resto/upgrade that never happened a few years ago, and they will fit perfectly here. The input cap is a small axial polyester 470nF, it fits well. I did consider buying the assembled item, as you say they are for nothing, however tales of fake bits here dissuaded me. In addition, I want to build the thing. Screwing some boards to a heat sink didn't have the same appeal. Happily I have a Farnell trade counter on this side of town, so I can at least get the bits. It all keeps me off the streets.