I forgot to say. Archimedes was killed by a Roman soldier when orders had been given to leave him alone. This says more than anything about stupidity and empire building. It is fitting that elephants should bring down that empire. Sicilly is a wonderful place. Malta also.
A quote from whoever. " The one thing that history teaches us is that history teaches us nothing". Engineers can better that. Old stuff done better can be better stuff. I see this on motorcars all the time. Not the posh things, bushes for example.
A quote from whoever. " The one thing that history teaches us is that history teaches us nothing". Engineers can better that. Old stuff done better can be better stuff. I see this on motorcars all the time. Not the posh things, bushes for example.
Nigel, I go along with that strongly ... my "bent" is to continually, steadily refine something that already exists, to "better the current" - this is something I have done all my life in all areas of interest, it's my "pleasure",
. Take a garden, for example - some people will look at what exists now; say, "I don't like it!", rip everything up, change the landscape entirely, buy in tonnes of top soil, buy a whole, "instant" garden of plants - and then walk away from it, it's no longer interesting. Not me, I look at the current arrangement of plants, form a picture of how the best parts of it can be made better, and over a period of time slowly make the garden evolve to a "superior", more satisfying place to be - the journey of making the changes is what gives me a buzz ...
Last edited:
I agree with Frank. Until recently, I was not aware how much the change of just one thing can bring about in terms of quality. I knew everything mattered, but the sheer magnitude of the a single component change I would not have believed possible untuil it happened to me. I refer to the pot change in my H/K 6550 integrated amp, when a cheap pot was replaced by a standard ALPS Blue.
Live through it just once and you are in it forever. That's why I can easily relate to Frank's efforts. Building up my Marantz' 170 DC power supply only confirms it.
Live through it just once and you are in it forever. That's why I can easily relate to Frank's efforts. Building up my Marantz' 170 DC power supply only confirms it.
I was just saying this about the Scott motorcycle of WW1 vintage to a friend. The engine with some modern stuff to give the sparks is very fast and not too bad on smoke ( two stroke ). I asked my friend why the bearings have a bearing plate ? Simple. The bearing plate is a five minute job to open. If a spare set were in the pocket they could be fitted in 10 minutes. If not the blue ones from one side mixed with the OK ones from the other side to get you home. Not bad hey?
One thing I beg. Please peope who make parts use paint I can read. Black on black is impossible. I have some BC327 with silver paint with black words. It is so easy to read. The face of BC327 is small enough to make it difficult. This way proves it to be easy.
One thing I beg. Please peope who make parts use paint I can read. Black on black is impossible. I have some BC327 with silver paint with black words. It is so easy to read. The face of BC327 is small enough to make it difficult. This way proves it to be easy.
Yes, some years ago in developing a preamp circuit, I compared two different RIAA capacitors of the exact same manufacturer, measured value, and voltage rating.
The only differences were the dielectrics, which were polystyrene and polypropylene. The latter was far better sonically in every way, even though I had expected the opposite.
Ray, I expected a change, which is why I made it. I wanted and expected better.
What bowled me over was not that there was a change,, but what it translated to in real life. Form a reasonable but unexciting integrated amp, with many virtues screaming that I should be getting more, it instantly transformed to a sound I have heard only a few times in my life, and for prices at least twen time (!!!) what this one cost.
Believe it or not, but for about two months, I couldn't believe what I was hearing, I kind of waited to wake up at any moment and confirm this was a dream only. But it wasn't. That's why I honestly admit that if anyone had told me this, I would not have belived it possible.
That's why I tend to keep Frank's side, firstly because I've just been there and experienced that, and secondly because I think the engineers here sometimes tend to be a little too bull headed.
For example, I mentioned several times that the pro sector of the industry keeps using now seemingly outdated MJ series of metal TO-3 power amplifier transistors for more than just getting more power out of fewer units, it's their sound. That was eliminated quickly enough by people pointing out that these were slow devices and were superceeded by newer and much faster ones. Since the tone was dismissive, I didn't ealborate, but Ihonestly feel their SOUND is generally better than that of most other modern devices. Yes, their Ft is around 6 MHz only, but they are by and large much less prone to oscillation than ten times faster Sanken devices, for example, and the fact that they can handle considerably more power than most plastic TO-247 packages should not be overlooked.
The acid test is to build the same circuit twice, once with modern devices, and again with MJ devces, you nly have to adapt the driver as their gain is lower - then read their respective THD and IM levels, and 1 will get you 10 that the MJ devices will do better, all esle being equal. How is that a bad thing?
I feel there's too much art for art's sake going on here, and agree with John when he says we should not lose sight of the goal. There's a zeros tace in THD and IM here.
What bowled me over was not that there was a change,, but what it translated to in real life. Form a reasonable but unexciting integrated amp, with many virtues screaming that I should be getting more, it instantly transformed to a sound I have heard only a few times in my life, and for prices at least twen time (!!!) what this one cost.
Believe it or not, but for about two months, I couldn't believe what I was hearing, I kind of waited to wake up at any moment and confirm this was a dream only. But it wasn't. That's why I honestly admit that if anyone had told me this, I would not have belived it possible.
That's why I tend to keep Frank's side, firstly because I've just been there and experienced that, and secondly because I think the engineers here sometimes tend to be a little too bull headed.
For example, I mentioned several times that the pro sector of the industry keeps using now seemingly outdated MJ series of metal TO-3 power amplifier transistors for more than just getting more power out of fewer units, it's their sound. That was eliminated quickly enough by people pointing out that these were slow devices and were superceeded by newer and much faster ones. Since the tone was dismissive, I didn't ealborate, but Ihonestly feel their SOUND is generally better than that of most other modern devices. Yes, their Ft is around 6 MHz only, but they are by and large much less prone to oscillation than ten times faster Sanken devices, for example, and the fact that they can handle considerably more power than most plastic TO-247 packages should not be overlooked.
The acid test is to build the same circuit twice, once with modern devices, and again with MJ devces, you nly have to adapt the driver as their gain is lower - then read their respective THD and IM levels, and 1 will get you 10 that the MJ devices will do better, all esle being equal. How is that a bad thing?
I feel there's too much art for art's sake going on here, and agree with John when he says we should not lose sight of the goal. There's a zeros tace in THD and IM here.
Last edited:
Guys, I've pulled almost 20 posts that are about politics, which is against the rules here. The instigator was someone who just got a vacation from here for exactly this violation. He'll be missing for a while, but please be warned that if any of that is continued here, others will be joining him in read-only mode.
The polystyrene cap caused the sound to be rather soft, veiled, textured, and diffuse.
By comparison, the polypropylene cap was substantially more transparent, and more extended in the highs.
These were both 630V rated, values matched to 0.1%, and with their outer foils oriented in the same way.
Demian, I am not clear at all as to why you call this a cascode.
To me I see a well implemented "standard" cascade of two differential stages. Coupling between the CCS's of the two pair, but nothing that denotes a cascode to me. Although Q4's drive is not defined in this clip.
I see the CCS of Q1 and Q2 being driven by the sum of the two diff pairs Q3 and Q5/6. Should this not (in the linear range) be constant? Only any errors in matching would be present here.
C13/14 look to me as part of the input protection scheme and not controlling the FET input C in any meaningful way.
What am I missing?
Thanks,
Alan
I would tend to suspect that it's the thermal behaviour of the TO3 package that's key ... once one gets normal technical competence in order then it's all the second order effects that become audible - and this can be a nightmare because very little of it is thoroughly documented, or easily testible. However, enough people can hear their impact, which makes it worthwhile pursuing the ridding a system of these artifacts - yes, that goal is all-important ...
Sorry that you trod on the banana twice, Dejan ...
Last edited:
Dejan, me also. TO3 might sound better in that the epoxy stops heat getting out and the TO3P footprint also. I have TIP3055 boiling away at 70C at the heat sink. That means above boiling at the collector. The epoxy seems cooler than the heatsink. Not so when using a 2N3055, it was hot. 4 degrees per watt and 2.5 ( TO3 ) case to heatsink when ideal which it never is. I seem to remember ? SOA is much worse than often suspected. I recently found out the FET SOA is based on tollerating 300 C briefly. Bipolar is far less. FET's still suffer local hot spots spoiling the party. These things must matter.
I made a very good cheap TO3P mount. 2.8 deg/watt heat sink ( Fischer ) . 3.5 mm hole deburred. Small rubber feet to the back. Heat compond. Wind in woodscrew to fix down. The rubber firmly mounts the device as it works like a spring. The feet allow air to pass under. I didn't have a PCB type so had to improvise. It seems a very easy way to try an idea. I doubt it took me 2 minutes to do.
I made a very good cheap TO3P mount. 2.8 deg/watt heat sink ( Fischer ) . 3.5 mm hole deburred. Small rubber feet to the back. Heat compond. Wind in woodscrew to fix down. The rubber firmly mounts the device as it works like a spring. The feet allow air to pass under. I didn't have a PCB type so had to improvise. It seems a very easy way to try an idea. I doubt it took me 2 minutes to do.
It seems I underestimated T03 and TO3P thermal transfer. They seem a bit optimistic to me. People do long chains of calculations. Seldom does that seem to relate to real life. I seem to always to need 50% more heatsink. My data came from T0220 voltage regulators that are many times worse.
This might also be a sound quality problem. One idea I had years ago was to cut the heads off of T03. An O-ring to seal the collector into a oil bath with radiator and pump. Although the collector makes less good contact this way the oil can get to the silicon slice. I suspect this will sound better. Cooking oil should be OK as a start. Everything would need to be very clean. It is just possible water could be used if very very clean ( Domestic filter element ? ) . 4.2Kj/kgK is better than 1.3Kj/kgK. Water has no equal.
If wanting a bit more the collector back could use the clamp as a heat shunt. It could have a pipe each side to help the heat shunt away. The emitter and base coming through the metal. Easy to do and I suspect worth 10% better heat transfer. There would be no heat transfer washer . NPN and PNP on isolated sections. If wanting to save money this could be 22mm/ 3/4 copper pipe. If could be folded in the middle and soldered where drilled. If the T03 stood 0.5 mm on it's O-ring it could be clamped down using the T03 holes to seal. The voltage rail to the pipe.
This might also be a sound quality problem. One idea I had years ago was to cut the heads off of T03. An O-ring to seal the collector into a oil bath with radiator and pump. Although the collector makes less good contact this way the oil can get to the silicon slice. I suspect this will sound better. Cooking oil should be OK as a start. Everything would need to be very clean. It is just possible water could be used if very very clean ( Domestic filter element ? ) . 4.2Kj/kgK is better than 1.3Kj/kgK. Water has no equal.
If wanting a bit more the collector back could use the clamp as a heat shunt. It could have a pipe each side to help the heat shunt away. The emitter and base coming through the metal. Easy to do and I suspect worth 10% better heat transfer. There would be no heat transfer washer . NPN and PNP on isolated sections. If wanting to save money this could be 22mm/ 3/4 copper pipe. If could be folded in the middle and soldered where drilled. If the T03 stood 0.5 mm on it's O-ring it could be clamped down using the T03 holes to seal. The voltage rail to the pipe.
Thermal camera....best tool for thermal engineering with real products, mimicking enclosures can be fun.
Have a look around this site, it also lists suitable fluids...
Iceotope is the home of cutting edge liquid cooling technology | Iceotope
These can do some funky stuff (as can bequest) had some cool heat pipe assemblies from them in the past...
Aavid Thermalloy - The World Leader in Thermal Technologies
Have a look around this site, it also lists suitable fluids...
Iceotope is the home of cutting edge liquid cooling technology | Iceotope
These can do some funky stuff (as can bequest) had some cool heat pipe assemblies from them in the past...
Aavid Thermalloy - The World Leader in Thermal Technologies
Does anyone know anything about Schottky diodes? SB540 or similar. In the past I tried some and was convinced the sound was worse. A friend said although not having the classical PN junction problems they do have nasty RF output . That seemed strange as how they work seems to say otherwise. I would use soft recovery diodes.This time
I can not afford the heat output and losses. That being so they need suppressing. I suppose 10 nF to the AC input of the bridge is not going to do it?
Aavid were is Swindon. They made me a batch of heatsinks once.
I can not afford the heat output and losses. That being so they need suppressing. I suppose 10 nF to the AC input of the bridge is not going to do it?
Aavid were is Swindon. They made me a batch of heatsinks once.
Opamp noise calculator
I entered my Lyra PU into this calculator and have a 83dB noise figure before amplifiaction.
I entered my Lyra PU into this calculator and have a 83dB noise figure before amplifiaction.
Maybe cascode is not the best term here. The circuit actively keeps the Vds of the FETs constant which reduces the cap modulation at the input of the FETs. This reduces the distortion at the input significantly when the source impedance is higher than 50 Ohms.
The measured distortion products of the amp circuit is below -140 dB at a gain of around 10X. That's a pretty good validation of the circuit.
- Status
- Not open for further replies.