Hello Nigel,
A little RR story, used to call at RR Mountsorrel Fabrications, they closed it, oooops! Sometime later I was walking around Earl Shilton and came across the two blokes who spun stuff at RR. RR forgot about the nacelles and nose cones (want one for your C47), recalled the two blokes, gave them the spinning machines and all the wooden chuck patterns. They set up in ES and RR had to send the quality guys in so they could have RR certification, they knew nothing about metal spinning, showed them how to do the paperwork and quickly left.
One of the spinning lathes would throw an 8ft blank and both of them together had to spin the steel over the wooden former with a roller ended metal 'pole' and do it in one go at 3,000 RPM. I stood well back while they did it and laughed at me.
I came across a dual spinning Taylor lathe in the Jewellery Quarter, a lady operated it to make posh silver trophy cups, I know where to find the date on the Taylor ones looked and cleaned it off a bit to reveal 1907. I asked her if I could feel the spindles, no movement. She used a steel bar with a rounded and polished end to it and what come off that machine just needed polishing.
Cheers
A little RR story, used to call at RR Mountsorrel Fabrications, they closed it, oooops! Sometime later I was walking around Earl Shilton and came across the two blokes who spun stuff at RR. RR forgot about the nacelles and nose cones (want one for your C47), recalled the two blokes, gave them the spinning machines and all the wooden chuck patterns. They set up in ES and RR had to send the quality guys in so they could have RR certification, they knew nothing about metal spinning, showed them how to do the paperwork and quickly left.
One of the spinning lathes would throw an 8ft blank and both of them together had to spin the steel over the wooden former with a roller ended metal 'pole' and do it in one go at 3,000 RPM. I stood well back while they did it and laughed at me.
I came across a dual spinning Taylor lathe in the Jewellery Quarter, a lady operated it to make posh silver trophy cups, I know where to find the date on the Taylor ones looked and cleaned it off a bit to reveal 1907. I asked her if I could feel the spindles, no movement. She used a steel bar with a rounded and polished end to it and what come off that machine just needed polishing.
Cheers
Worra nerd
Sorry all,
Forgot to delete the resistor when I added the trim pot
Cheers
Sorry all,
Forgot to delete the resistor when I added the trim pot
An externally hosted image should be here but it was not working when we last tested it.
Cheers
I knew Doreen Jordan who spun the Jordan metal cone speakers. Not a sound I especially like. My Eminence 12 LTA stuff similar collocation.
On that subject even on 78s I could easily hear the tweeter. A wider window even if surface noise. The beaming of a Wizzer cone is distracting. The Tannoy's despite a sophisticated tweeter are similar.
If making baffle speakers you have to unlearn everything you believe you know. Luckily others have made the mistakes for you. Follow what they recommend. Big drivers in open air can be remarkably good. Some colouration is typical. Listening fatigue can be lower than box speakers. There is a new Fane drive unit doing the rounds. I am dubious about the Fane and prefer what I did. The tweeter I used was high efficiency as it's prime quality. A really good tweeter would have been a mistake. The maths need to work. Because I crossed at 6kHz many distortions of a cheaper tweeter avoided. At my age it shouldn't matter. It does because the speed of how we hear remains fast even when the frequency range reduces. 78s carry information the shellac due to grain size couldn't give, the leading edge still there. Decca ffrr 78s give 32 kHz if Vinyl pressings. This is why some hear improvements in the JLH that should be impossible.Its more about square waves.
On that subject even on 78s I could easily hear the tweeter. A wider window even if surface noise. The beaming of a Wizzer cone is distracting. The Tannoy's despite a sophisticated tweeter are similar.
If making baffle speakers you have to unlearn everything you believe you know. Luckily others have made the mistakes for you. Follow what they recommend. Big drivers in open air can be remarkably good. Some colouration is typical. Listening fatigue can be lower than box speakers. There is a new Fane drive unit doing the rounds. I am dubious about the Fane and prefer what I did. The tweeter I used was high efficiency as it's prime quality. A really good tweeter would have been a mistake. The maths need to work. Because I crossed at 6kHz many distortions of a cheaper tweeter avoided. At my age it shouldn't matter. It does because the speed of how we hear remains fast even when the frequency range reduces. 78s carry information the shellac due to grain size couldn't give, the leading edge still there. Decca ffrr 78s give 32 kHz if Vinyl pressings. This is why some hear improvements in the JLH that should be impossible.Its more about square waves.
Hello Nigel,
Some of those old recordings sound great even now e.g.
Bessie Smith with Clarence Williams, acoustic recording 'Oh daddy blues' c.1923
Geoffrey Smith on JRR introduced this as a 3 minute masterpiece
Billie Holiday, with Teddy Wilson, Lester Young etc, 1937 'Me myself and I'
And ended it emotionally with 'Worra rekerd' (nearest I can get to his accent)
[edit] and Beethoven's 7th Berlin Philharmonic, Wilhelm Furtwängler, 1943 live recording
The epitome of dance; in the USA, Martha Stewart and Merce Cunningham, and of course Isadora Duncan and our very own Margaret Morris
Cheers
Some of those old recordings sound great even now e.g.
Bessie Smith with Clarence Williams, acoustic recording 'Oh daddy blues' c.1923
Geoffrey Smith on JRR introduced this as a 3 minute masterpiece
Billie Holiday, with Teddy Wilson, Lester Young etc, 1937 'Me myself and I'
And ended it emotionally with 'Worra rekerd' (nearest I can get to his accent)
[edit] and Beethoven's 7th Berlin Philharmonic, Wilhelm Furtwängler, 1943 live recording
The epitome of dance; in the USA, Martha Stewart and Merce Cunningham, and of course Isadora Duncan and our very own Margaret Morris
Cheers
Last edited:
78's often sound better than modern recordings. Benny Goodman was also a great sound engineer. Somethings he did were slightly dangerous like climbing a box to sing. Benny made everything work with one microphone. Stereo can be as simple. The time relationships remain correct. Lets be critical here. Only this way is high fidelity. Many 78's were using close microphones. I still hear the room. 1938 was about the peak. Pink Floyd is different. The sound is the sound. You will find what is right for 78s is right for PF. In fact early Floyd is very 1938 in style. The sound of a dripping tap in space. Modern styles are like musical mincemeat.
This recording will sound wrong on most hi fi's. It not the recording. It's like not having the right glasses to see. Listen carefully to the silent bits. The colourations are often microphone placement. This microphone more or less is still made. It's very expensive. Piano is hard to record this way. It can sound like it is underwater. Bias for magnetic recordings a recent addition as here. This machine became Ampex. This recording shows what of the 78 sound is microphones as it is similar. The second movement my favourite of this piece. The BBC used this template for many years. So called Blumlein Stereo by some. They somehow made it bland.
YouTube
This recording will sound wrong on most hi fi's. It not the recording. It's like not having the right glasses to see. Listen carefully to the silent bits. The colourations are often microphone placement. This microphone more or less is still made. It's very expensive. Piano is hard to record this way. It can sound like it is underwater. Bias for magnetic recordings a recent addition as here. This machine became Ampex. This recording shows what of the 78 sound is microphones as it is similar. The second movement my favourite of this piece. The BBC used this template for many years. So called Blumlein Stereo by some. They somehow made it bland.
YouTube
Hello all
Blank tag board layout
I've left a big gap between the rows so the components can be nominated.
Cheers
I'll have a go at populating it for the Zerozone layout below, without the led bit
Blank tag board layout
An externally hosted image should be here but it was not working when we last tested it.
I've left a big gap between the rows so the components can be nominated.
Cheers
I'll have a go at populating it for the Zerozone layout below, without the led bit
An externally hosted image should be here but it was not working when we last tested it.
Whilst you can get away with almost anything for the semis in JLH'69, I wouldn't stray too far from the types originally specified. So I'm wondering how a general purpose amplifier and switching type like 2N2907 for the input stage on the Zerozone schematic, came about. I think you would be better off with a low noise type, i.e. at least as low noise and high gain as typical 2N3906. Perhaps that's only a nominal type used for drafting purposes but original type markings seem to be more typical of Chinese kits.
My parts collection is of mainly European BC prefix types so I might use BC559, 560 or even the their metal can versions if the finished boards need to look olde worlde but that's just what I have on hand. There are plenty of other suitable types to play about with among US and Japanese types.
My parts collection is of mainly European BC prefix types so I might use BC559, 560 or even the their metal can versions if the finished boards need to look olde worlde but that's just what I have on hand. There are plenty of other suitable types to play about with among US and Japanese types.
Last edited:
There are switching types that were used as they can be ultra low noise. 2N4401/3 for example. Power transistors also. Rbb- or base spreading resistance seldom is quoted. I like BC 327/337. BC327 is 0.6nV/✓Hz which is very close to the absolute limit. BC327 is high current and gain. It's voltage about the same as other good devices. In the day small power transistors in T0126 packages were used. I suspect BD136 would qualify . It is likely that JFETs would suit the input of the JLH better due to the very sensible input.inpedance he used. I seem to remember rbb- of about 1 ohm was the ideal. 20 more typical. JFETS like in Hiraga Le Monstre are the low noise types. I have no experience of substituting them in this circuit. Their amplification curves are slightly more valve like. They can if forced work in enhancement mode like a transistor. I doubt that is ideal. They are depletion devices like valves. Op amps have very simple biasing which isn't possible here. Tl071 for example.
https://global.oup.com/us/companion...icrocircuits/students/bjt/2N2907-motorola.pdf
https://www.farnell.com/datasheets/727135.pdf
The base spreading resistance of BC560 is not hyper low. However it is very low noise. 2N2907 also. Switching means fast. The very best transistors we generally had were TV types often used for SE class A gun drivers. BF720/721. Surprisingly the linearity was good enough. In the old days we had to accept ideas that circulated. These days we have PDF. 2SA1085 if genuine is a very good device. It was a microphone amplifier of higher dissipation in TO92 long case of circa 0.75 watts. Very high current for a non switching type. MPSA92/42 are sometimes a good choice. 2N2907 has a higher input capacitance which might be OK or even helpful. I suspect this device was used as it has the 1960s look. BC109 is a supurb device.
https://www.farnell.com/datasheets/727135.pdf
The base spreading resistance of BC560 is not hyper low. However it is very low noise. 2N2907 also. Switching means fast. The very best transistors we generally had were TV types often used for SE class A gun drivers. BF720/721. Surprisingly the linearity was good enough. In the old days we had to accept ideas that circulated. These days we have PDF. 2SA1085 if genuine is a very good device. It was a microphone amplifier of higher dissipation in TO92 long case of circa 0.75 watts. Very high current for a non switching type. MPSA92/42 are sometimes a good choice. 2N2907 has a higher input capacitance which might be OK or even helpful. I suspect this device was used as it has the 1960s look. BC109 is a supurb device.
R A Penfold saw great virtue in circuit designs on paper being like the real thing as JLH. His little books are gems of common sense. He often refers to others and I know he meant JLH.
One way to do point to point is draw it real size. I use brass No4 wood screws. Plywood is good.
In the circles were I have friends point to point is considered the best of the best. Overlap the component wires where possible rather than going through the brass..
Point to point allows semi 3D layouts. Sometimes a wire is better than a PCB track as it can be bent to work better on the oscilloscope. Thicker wire can be better.
One way to do point to point is draw it real size. I use brass No4 wood screws. Plywood is good.
In the circles were I have friends point to point is considered the best of the best. Overlap the component wires where possible rather than going through the brass..
Point to point allows semi 3D layouts. Sometimes a wire is better than a PCB track as it can be bent to work better on the oscilloscope. Thicker wire can be better.
Hello Nigel,
The screws sound interesting I can see how the schematic could be used directly. I think about 200 would be enough for the power supply, pre-amp and main boards. Only £10.40 for the 200: Solid Brass Wood Screws Raised Countersunk Slotted Head No.4, No.6, No.8 Gauge | eBay
Cheers
The screws sound interesting I can see how the schematic could be used directly. I think about 200 would be enough for the power supply, pre-amp and main boards. Only £10.40 for the 200: Solid Brass Wood Screws Raised Countersunk Slotted Head No.4, No.6, No.8 Gauge | eBay
Cheers
Just before I moved house I made a version of the 1969 circuit. The outputs Indian 2N3055 that I suspect were rather good. BC327/337. I had to make a very low distortion oscillator from Elektor magazine to measure it. 0.03% distortion. 6 watts. It was point to point. I made a serious design error using BC337 which by virtue of how high the 2N3055 gain was I got away with. BC337 being a bit small. The transistor never got hot which surprised me. I wish Andrew T was still with us.I would have liked him to investigate how this driver worked. I used point to point and 33pf collector to bsse on the driver. Andrew shared he didn't quite understand the design if I remember correctly. I think JLH is a superb example of when negative feedback works. It really shouldn't work. Where possible I avoid negative feedback. Not here although even than can be tweaked. In my oppion very low open loop gain and sensible use of parts makes for a miracle. Even the damping factor is better than seems possible.
Hello Nigel,
You metioned plywood
I make my boxes from the cheapest ply 1220 x 610 for £6.42
Really easy stuff to work with.
Cheers
You metioned plywood
An externally hosted image should be here but it was not working when we last tested it.
I make my boxes from the cheapest ply 1220 x 610 for £6.42
Really easy stuff to work with.
Cheers
Let's just consider the original PNP input device. We could go on ad infinitum about reworking the design and using other devices but most folk are going to want or need to start out at a reference point - i.e. the original JLH'69 design with 2N3906 that came in a TO92 package.
It was originally a Motorola transistor ca. 1967 but the spec. was not the same as currently available product. 'same goes for 2N2907 and 2N2907A which are also somewhat different between brands and types.
I suspect that if you fit whatever 2N2907 or 2N1711 etc. marked transistors come with the kit, you won't be getting those types anyway. I don't trust the low price kits that come with expensive looking new components masquerading as NOS. The last time that I picked up one of the TO18 can transistors, its ink marking transferred to my finger but perhaps others will appreciate such comments a little better when they receive the currently available kits rather than building from scratch and the junkbox. My junkbox probably dates back as far as 1970 and I doubt many others will be using those vintage components now either.
It was originally a Motorola transistor ca. 1967 but the spec. was not the same as currently available product. 'same goes for 2N2907 and 2N2907A which are also somewhat different between brands and types.
I suspect that if you fit whatever 2N2907 or 2N1711 etc. marked transistors come with the kit, you won't be getting those types anyway. I don't trust the low price kits that come with expensive looking new components masquerading as NOS. The last time that I picked up one of the TO18 can transistors, its ink marking transferred to my finger but perhaps others will appreciate such comments a little better when they receive the currently available kits rather than building from scratch and the junkbox. My junkbox probably dates back as far as 1970 and I doubt many others will be using those vintage components now either.
I think the plywood is a great economy and can look pretty good too but I'd also think about the much greater heat you'll have to deal with if you choose to use plywood for a class A amplifier enclosure, jemraid. Even a 10-15W JLH amplifier pours out the heat and you need to keep it away from the wood otherwise that dries out and distorts , turning it into a brittle, delaminating mess. Consider large heatsinks (<0.5 °K/W) and generous metal mesh vents top and bottom.
When fitting transistors Vero pins help. Changing a transtor must be the more interesting easy thing to do. Vero pins make it simple. The fake types often better.The real ones a bit chunky. Try a power transistor for fun.A genuine BD140/136.
Plywood. Look at my baffles. Standard 2 x 4 x1/2 inch ply and MDF glued together. Buckets of water to hold them together. PVA glue. Notice the concave shape. Accident of the bonding. Cut using Elu router and tramel bar. The positioning is to use the floor as part of how it works.
The BBC seems to be trolling our government. What of other countries. It's not nice to troll and now is not the moment. It is not good to question the use of the conditional tence. The Welsh minister in the nicest way told them off. Bruno the French minister was excellent and impressed me. I am not great in French. He made it easy.
Plywood. Look at my baffles. Standard 2 x 4 x1/2 inch ply and MDF glued together. Buckets of water to hold them together. PVA glue. Notice the concave shape. Accident of the bonding. Cut using Elu router and tramel bar. The positioning is to use the floor as part of how it works.
The BBC seems to be trolling our government. What of other countries. It's not nice to troll and now is not the moment. It is not good to question the use of the conditional tence. The Welsh minister in the nicest way told them off. Bruno the French minister was excellent and impressed me. I am not great in French. He made it easy.
Hello Ian,
Thanks for your message and for your interest in my humble project.
I will be mounting the heat sinks on the sides of my box, they can be seen here;
https://www.diyaudio.com/forums/solid-state/3075-jlh-10-watt-class-amplifier-610.html#post6137952
I agree that heat will warp the wood and will also fit some diagonal braces like the Romans did to cross the Danube.
Hello Nigel,
I did try a layout for those tag boards but the amount of hook up wires defeated me.
I agree about the Veropins I'm going to use lots of them I have in mind another so called 'better' set of transistors and if the JLH doesn't exceed my 3886 I'll buy some.
Cheers
Thanks for your message and for your interest in my humble project.
I will be mounting the heat sinks on the sides of my box, they can be seen here;
https://www.diyaudio.com/forums/solid-state/3075-jlh-10-watt-class-amplifier-610.html#post6137952
I agree that heat will warp the wood and will also fit some diagonal braces like the Romans did to cross the Danube.
Hello Nigel,
I did try a layout for those tag boards but the amount of hook up wires defeated me.
I agree about the Veropins I'm going to use lots of them I have in mind another so called 'better' set of transistors and if the JLH doesn't exceed my 3886 I'll buy some.
Cheers
Try any PNP. The famous Lentek MC pre amp used power transistors. It was supurb.
Try also three transistors in parallel. Match them best you can. I agree with Douglas Self. That despite theory that can work. It still is a single device in how it works. If one takes more current it doesn't really matter A double BC 337-40 VAS could work. It still has more gain than typical devices. Doubtless it wouldn't need compensation.
Try also three transistors in parallel. Match them best you can. I agree with Douglas Self. That despite theory that can work. It still is a single device in how it works. If one takes more current it doesn't really matter A double BC 337-40 VAS could work. It still has more gain than typical devices. Doubtless it wouldn't need compensation.
Regarding suitable transistors, it is hard to see that a BC557 or BC558 would not perform as well as a 2N3906. The fT is similar, noise figure similar.
For the driver, the 2N1613 was a rather standard device at the time. A BD139 should work, though I have simulated with 2N3019 (TO-39, still available from Farnell, but not sure who actually manufactures the Multicomp device) with better (sim) results. The old devices tended to have higher Cjc, so modern ones might need a small compensation (c.22pF).
I like the idea of using 2 or 3 BC337's, but I would use small emitter resistors to ensure current sharing, which obvs. changes the design a little.
I have used 2SC5200's without trouble, but MJL3281A's needed a global 33pF compensation. 2N3055's will work "as it says on the tin" but the modern epi is only rated at 2.5MHz fT, while the original (MJ480's) were 4MHz. Modern 4MHz devices tend to be 150W - 250W and may have large input capacitances which may need (ideally) a driver, but MJ21194 simulates OK at 20kHz.
For the driver, the 2N1613 was a rather standard device at the time. A BD139 should work, though I have simulated with 2N3019 (TO-39, still available from Farnell, but not sure who actually manufactures the Multicomp device) with better (sim) results. The old devices tended to have higher Cjc, so modern ones might need a small compensation (c.22pF).
I like the idea of using 2 or 3 BC337's, but I would use small emitter resistors to ensure current sharing, which obvs. changes the design a little.
I have used 2SC5200's without trouble, but MJL3281A's needed a global 33pF compensation. 2N3055's will work "as it says on the tin" but the modern epi is only rated at 2.5MHz fT, while the original (MJ480's) were 4MHz. Modern 4MHz devices tend to be 150W - 250W and may have large input capacitances which may need (ideally) a driver, but MJ21194 simulates OK at 20kHz.