What a beautiful amp Mr JLH. Who could imagine a man worse for taking parts out of an amplifiers than me could make that? My silly amp I show is JLH with MOS FET's . It worked but was one step too far. JLH even got the JLH 10 watt amp to have DC coupled speakers without LTP input (me also).
What JLH might have had was what I have, a team of engineers to help me. He might have guided the design. After years of doing it all myself I have that. Problem is many projects don't get finished because I am getting too old to do it all. In the past I did. I can't be told off because I always do 110%. Real life needs 130%. To be honest where I do that it is a coded message to my boss that I don't agree. Not sure he would ever understand that? I never say no.
I am semi retired, yet work harder than ever. My divorce and me not having money when young means I work on. I guess you all know I wouldn't have it differently.
I recently for no salary took control of my old company and re did the pricing structures. Money has returned to the company when it was badly in the black a few weeks ago. That is the 130% of doing work others are paid to do. I did it so all my hard work isn't lost in that company. It took me two hours. The letter of apology to the dealers one of my greatest moments. It simply said the banking crisis hurt us and we did what we had to to survive. Then said we have a new pricing structure and lower prices negotiated on the basis of the hoped for better sales. It worked like a charm as none of it was untrue. Now guys, I never trained for that. The lady who has that job hates me. My name is never mentioned. It is because I can not conceal how I feel typing some invoices is not really financial management. She sees it differently. She almost pulled the company down and is oblivious to that. She has the shock of her life coming as I have negotiated that our subcontractor makes a cheaper product for the old boss. It will be sold by both companies and respect the origial as the designers (that's me although it won't be stated). It is a horrible risk and might back fire. It will take honorable working on all sides. It might be OK as the situation was so bad it needed something. I will get a very small percentage on each sale which was not even in my head when I did it. I might have to lend the old company money to do it. I always respect Percy Wilson in that story as he was with a Mr Edgar German the co-inventor in 1962/68. I walk past Percy's house often. The Mr E G is as best as I know the correct name as he preferred not to be named. Percy had been writing about this since the 1920's. Another gentleman I care not to name thinks they were incidental. If so Mr Einstein also. This has been the easiest thing I ever did in my life and yet wasn't able to be done by anyone else it seems.
What JLH might have had was what I have, a team of engineers to help me. He might have guided the design. After years of doing it all myself I have that. Problem is many projects don't get finished because I am getting too old to do it all. In the past I did. I can't be told off because I always do 110%. Real life needs 130%. To be honest where I do that it is a coded message to my boss that I don't agree. Not sure he would ever understand that? I never say no.
I am semi retired, yet work harder than ever. My divorce and me not having money when young means I work on. I guess you all know I wouldn't have it differently.
I recently for no salary took control of my old company and re did the pricing structures. Money has returned to the company when it was badly in the black a few weeks ago. That is the 130% of doing work others are paid to do. I did it so all my hard work isn't lost in that company. It took me two hours. The letter of apology to the dealers one of my greatest moments. It simply said the banking crisis hurt us and we did what we had to to survive. Then said we have a new pricing structure and lower prices negotiated on the basis of the hoped for better sales. It worked like a charm as none of it was untrue. Now guys, I never trained for that. The lady who has that job hates me. My name is never mentioned. It is because I can not conceal how I feel typing some invoices is not really financial management. She sees it differently. She almost pulled the company down and is oblivious to that. She has the shock of her life coming as I have negotiated that our subcontractor makes a cheaper product for the old boss. It will be sold by both companies and respect the origial as the designers (that's me although it won't be stated). It is a horrible risk and might back fire. It will take honorable working on all sides. It might be OK as the situation was so bad it needed something. I will get a very small percentage on each sale which was not even in my head when I did it. I might have to lend the old company money to do it. I always respect Percy Wilson in that story as he was with a Mr Edgar German the co-inventor in 1962/68. I walk past Percy's house often. The Mr E G is as best as I know the correct name as he preferred not to be named. Percy had been writing about this since the 1920's. Another gentleman I care not to name thinks they were incidental. If so Mr Einstein also. This has been the easiest thing I ever did in my life and yet wasn't able to be done by anyone else it seems.
Last edited:
That saved me much money. Note NP electrolytic on my amps.
The effects at - 150 dB !!!!!! That must be the ghost in the machine?
The Sony effects I mentioned were - 120 dB. That was vibration in amps.
Good reading, Jan.
This corresponds with my own views and practice rather well.
Now I know I wasn't just imagining things.
Just one thought - we often see audiophiles condaming ceramic caps. While this may be deserved by the El Cheapo discs, there also exist multilayer ceramic caps (in my case, from Siemens) which are among the best there are, but of course, do cost several times more than the usual ceramic fare.
This corresponds with my own views and practice rather well.
Now I know I wasn't just imagining things.
Just one thought - we often see audiophiles condaming ceramic caps. While this may be deserved by the El Cheapo discs, there also exist multilayer ceramic caps (in my case, from Siemens) which are among the best there are, but of course, do cost several times more than the usual ceramic fare.
cap tests -- limitations -
It should be noted --- that the assumption for the tests are that coupling capacitors are not properly used if there is a voltage developed across them. Such coupling caps need to be sized so that at the lowest freqs there will be no voltage drop across the cap. Then it is measured for harmonic and IM.
That is all fine and good. HOWEVER, there are a lot of places besides coupling cap applications where a large voltage WILL be dropped across the cap - deliberately; Such places are RIAA, any EQ, cross-overs, filters and tone control etc etc. There the distortion will be much higher than typically shown.
Still need to use the best dielectric film cap types.
THx-RNMarsh
It should be noted --- that the assumption for the tests are that coupling capacitors are not properly used if there is a voltage developed across them. Such coupling caps need to be sized so that at the lowest freqs there will be no voltage drop across the cap. Then it is measured for harmonic and IM.
That is all fine and good. HOWEVER, there are a lot of places besides coupling cap applications where a large voltage WILL be dropped across the cap - deliberately; Such places are RIAA, any EQ, cross-overs, filters and tone control etc etc. There the distortion will be much higher than typically shown.
Still need to use the best dielectric film cap types.
THx-RNMarsh
Last edited:
A few distortion measurements of popular caps with the newest Audio Precision should be all that is needed to get a sense of proportion about the level of distortion when say 1/2 voltage is dropped is across cap. Anyone can do this for themselves if they have the test equipment.
Dont need C.B to do it for you/us. Could even use a sound card and Victors osc and a notch filter approach. I would start with a bipolar used in speaker cross-overs. And, values used in EQ circuits/filters.
THx-RNMarsh
Dont need C.B to do it for you/us. Could even use a sound card and Victors osc and a notch filter approach. I would start with a bipolar used in speaker cross-overs. And, values used in EQ circuits/filters.
THx-RNMarsh
Last edited:
I remember my brother saying not to upgrade the disgusting looking ceramics on the VAS of my Hitachi amp. My intention to fit polystyrene. My brothers point was the wire to foil corrodes with time and then the amplifier destroys itself and the speakers. My brother sent me that magazine when it came out. The caps fitted were 50 for $1 and still NPO high grade. Often made for high stability tuned circuits. Notice NPO on my diagram. If you look very carefully NPO/COG = 30ppm. With a 50 ppm resistor they often form good stable oscillators. I had a Sony clock radio running on one to about 30 seconds a day,, a turntable PSU. Simon said silver-mica would be no better. Conversely I think I have seen polystyrene in Rotel doing this job and never remember it being a problem. RA 931 recently was to the best of my memory. Really nice amp and suits Dynaco A25's. They are my shame the rest references. Far better than the spec suggests.
Now a bit of trivia which may just about be correct. I had a service manual for BR locomotive type 47. Consumption about 0.8 mile per UK gallon. One only needs 100 passengers to say not bad. About a 400 mile + range at 90 MPH. Horse power about 2700. Next the English Electric shunter at 2 MPG and 600 horse (top speed 15 MPH). So I looked to the new Formula 1 cars which are 600 horse + 160 electric ( circa 800 as not exactly known) . 100 kg fuel at 0.711 kg/litre = 140 L = 30.94 UKG . Silverstone 3.66 miles x 52 = 190.4 miles. 6.2 MPG UK ( 5.12 US ). 46 L/100km . The laps speed time was typically 98 seconds. That relates to > 130 MPH average ( 215 km). As power rises as a cube that means to average 70 MPH as on UK motorway we need 8 times less fuel if all things are equal (they never are). 50 UK MPG (41 US) or 5.75 L/100 km. No petrol car I know of except Honda Jazz comes close. Now the killer blow. The cd of the best cars is about 0.27 and the worse 0.4, F1 is typical 1.2 as it has deliberate down-force. That fuel consumption is impressive. It is better than I state as none of the cars ran out. The EE shunter is a disgrace. I weigh slightly more than the fuel allowance for F1. Locomotive horses will be different as an FI engine would struggle to work the generators that drive the wheels. RPM of type 47 is 1150 max from memory with 950 RPM tick-over. The whistle you hear is the turbo charger. Hence the complex star delta motor systems to replace a gearbox. Soon all cars will benefit from this. My dream is cheap four wheel drive as a bonus and engine direct to wheels where possible. Myself I would have the engine rear mounted with electric front wheel drive. We might have a rear electric motor to help at low speed in the snow.
Now a bit of trivia which may just about be correct. I had a service manual for BR locomotive type 47. Consumption about 0.8 mile per UK gallon. One only needs 100 passengers to say not bad. About a 400 mile + range at 90 MPH. Horse power about 2700. Next the English Electric shunter at 2 MPG and 600 horse (top speed 15 MPH). So I looked to the new Formula 1 cars which are 600 horse + 160 electric ( circa 800 as not exactly known) . 100 kg fuel at 0.711 kg/litre = 140 L = 30.94 UKG . Silverstone 3.66 miles x 52 = 190.4 miles. 6.2 MPG UK ( 5.12 US ). 46 L/100km . The laps speed time was typically 98 seconds. That relates to > 130 MPH average ( 215 km). As power rises as a cube that means to average 70 MPH as on UK motorway we need 8 times less fuel if all things are equal (they never are). 50 UK MPG (41 US) or 5.75 L/100 km. No petrol car I know of except Honda Jazz comes close. Now the killer blow. The cd of the best cars is about 0.27 and the worse 0.4, F1 is typical 1.2 as it has deliberate down-force. That fuel consumption is impressive. It is better than I state as none of the cars ran out. The EE shunter is a disgrace. I weigh slightly more than the fuel allowance for F1. Locomotive horses will be different as an FI engine would struggle to work the generators that drive the wheels. RPM of type 47 is 1150 max from memory with 950 RPM tick-over. The whistle you hear is the turbo charger. Hence the complex star delta motor systems to replace a gearbox. Soon all cars will benefit from this. My dream is cheap four wheel drive as a bonus and engine direct to wheels where possible. Myself I would have the engine rear mounted with electric front wheel drive. We might have a rear electric motor to help at low speed in the snow.
There are many caps not intended for audio work like X2 polypropylene that are good. They sound good because what causes failure demands high standards of construction with sound byproduct. As they don't have idiot customers they are sold honestly.
I use the old skeleton form 100 to 400 V polyester in crossovers. Dirt cheap and good enough.
I prefer Panasonic NP caps for feedback loops. Said to better Black Gate if distortion means anything. Never polarize any cap if you can help it. Output cap or DC blocking cap no choice. PSU is different. Using a higher gain factor on amps keeps a NP away from critical 0.4V polarizing voltage. Although the cap doesn't mind your ears do. That dull restricted sound is that happening.
I use the old skeleton form 100 to 400 V polyester in crossovers. Dirt cheap and good enough.
I prefer Panasonic NP caps for feedback loops. Said to better Black Gate if distortion means anything. Never polarize any cap if you can help it. Output cap or DC blocking cap no choice. PSU is different. Using a higher gain factor on amps keeps a NP away from critical 0.4V polarizing voltage. Although the cap doesn't mind your ears do. That dull restricted sound is that happening.
Richard, a short while ago you commented on MOSFETs as being faster than BJTs regarding Ton, Tstore and Toff. You mentioned that BJTs need microseconds, while MOSFETs needed nanoseconds.
While I agree with you, I'd like to say, just for the record, that the above is not quite true. The somewhat strange thing is that we have had available some damn fast BJTs, quite comparable and in some specs, even faster than the current crop of fast output devices from Sanken, Toshiba, NEC et a., for over 30 years now.
I suspect they never made it outside of Europe because they were usually limited in voltages, typically up to 100V. Their general data sheets never failed to mention 120 and 140V versions, but I have never seen them anywhere (e.g. for BD 249/250, and many others).
I have attached the data sheet for one such device, BD 745, in this instance even made by a US company, Bournes. Take a look at its speed specs; not MOSFET class to be sure, but way faster than the vast majority of its contemporaries.
While I agree with you, I'd like to say, just for the record, that the above is not quite true. The somewhat strange thing is that we have had available some damn fast BJTs, quite comparable and in some specs, even faster than the current crop of fast output devices from Sanken, Toshiba, NEC et a., for over 30 years now.
I suspect they never made it outside of Europe because they were usually limited in voltages, typically up to 100V. Their general data sheets never failed to mention 120 and 140V versions, but I have never seen them anywhere (e.g. for BD 249/250, and many others).
I have attached the data sheet for one such device, BD 745, in this instance even made by a US company, Bournes. Take a look at its speed specs; not MOSFET class to be sure, but way faster than the vast majority of its contemporaries.
It's the switch off that counts. 100 nS on and 60 nS off Exicon. With bipolar the charge is slow to discharge. That is the reason for over-sized drivers and taking the basses down as far as you can below or to the center voltage.
The Fetlington has 0R1 and nice 100 mA across it. The BE is almost completely shunted. How can that be made quicker? Even that spike should be fast.
The switching speed is of zero importance except it screws up the feedback loop. As Bob Stewart said we mostly use a feedback amp to create a constant voltage source. How unlike the ideas often used a real amplifier is? Dejan that actually is a compliment to you.
Quad 405 is mostly a feedback amp. The feed-forward allows terribly slow transistor to work. Some as low as 300 kHz full power. Absolutely no difference or advantage in fitting faster ones. The technical reviewer stated in that alone current dumping feed-forward is a real solution. Up to that point he doubted it.
So we have feedback to control speakers and fast transistors to get a safety margin?
Then we have distortion reduced to absurdity mostly because we can.
The Fetlington has 0R1 and nice 100 mA across it. The BE is almost completely shunted. How can that be made quicker? Even that spike should be fast.
The switching speed is of zero importance except it screws up the feedback loop. As Bob Stewart said we mostly use a feedback amp to create a constant voltage source. How unlike the ideas often used a real amplifier is? Dejan that actually is a compliment to you.
Quad 405 is mostly a feedback amp. The feed-forward allows terribly slow transistor to work. Some as low as 300 kHz full power. Absolutely no difference or advantage in fitting faster ones. The technical reviewer stated in that alone current dumping feed-forward is a real solution. Up to that point he doubted it.
So we have feedback to control speakers and fast transistors to get a safety margin?
Then we have distortion reduced to absurdity mostly because we can.
That BD 745 is an excellent device and good price. Hard to say if compliment is as good. Sure it still rates highly. That with Exicon 10N/P20 would look good as Fetlington. If I understand it the switch off speed of a bipolar is theoretical when the FET is actual. Circuits that allow the bipolar to work correctly are not always the ones used.
BD746 Datasheet - Bourns | DatasheetLib.com
BD746 Datasheet - Bourns | DatasheetLib.com
And for even more "oomph!", choose BD 249/259 C. C is again for 100V, the original TI Power Device Data book mentions D for 120V and E for 140V, but I have never seen them offered anywhere.
Ton, Tstore and Toff times are similar to BD 745, but RMS current is quoted as 25A and peak current as 40A!
They stem from the same family as TI's TIP 35/36, but differ practically only by significantly shorter T times and a slightly higher price. This applies to TI's versions, no idea what ON Semi's versions are like.
In my view, BD 249/250 C are the unsung heroes of audio, provided you don't use more than +/- 40V PSU lines and use enough pairs for what you want. Unfortunately, they were not used often enough. Some Grudig models used them, most notably V 5000, and some UK models used TIP 35/36 C, notably Magnum. Both packed a punch, and especially the Grundig was well worth having, if reasobalbe power outputs (70/120W into 8/4 Ohms) was enough for you.
Ton, Tstore and Toff times are similar to BD 745, but RMS current is quoted as 25A and peak current as 40A!
They stem from the same family as TI's TIP 35/36, but differ practically only by significantly shorter T times and a slightly higher price. This applies to TI's versions, no idea what ON Semi's versions are like.
In my view, BD 249/250 C are the unsung heroes of audio, provided you don't use more than +/- 40V PSU lines and use enough pairs for what you want. Unfortunately, they were not used often enough. Some Grudig models used them, most notably V 5000, and some UK models used TIP 35/36 C, notably Magnum. Both packed a punch, and especially the Grundig was well worth having, if reasobalbe power outputs (70/120W into 8/4 Ohms) was enough for you.
Road map as usual. Add anything you think useful. As far as I know these are the topologies that switch Bipolars off fastest. The second has the nice idea of class A till the last moment. Instinct says it should be better. My feeling is the EF output stage has to be the better Bipolar stage for this reason alone. My Inca Talon was the first type.
The Fetlington shown as this is more or less all there is to it and worth showing for it's simplicity. I never knew how L FET's are drawn so no idea if right. Take letters as what you should know( N,P ,drain). It will be killer powerful. Even gain will be good. I suspect 90% as opposed to 78% for MOSFET alone. Bipolar EF about 93%.
About JL Hood . I forgot. His later designs were more complex. It must have been when transistors were expensive he designed the JLH 10 watt class A. It must be with the Hiraga Le Monstre the most simple respected amp.
Not PNP BD 746 ?
All true, but the reality is, nobody does it, and does a decent write-up, so far
Jan
I have done it. And sure others have also. But, we are not all interested in writing about it. I think it is enough to tell DIY'ers just to use this or that dielectric for best results and not write a book on it. No new tests have changed the basic dictum regarding polar and film.
It is also questionable that the results below -100dB are accurate with his test. The non-linearity of the ADC is poor there and small level changes will cause the harmonics to be shown in very different levels with respect to one another. Thats why you have to pay a small fortune to get accurate measurements below -100dB.
But, that is just an interesting side issue.... accurate harmonic measurements at very low levels.
Suggest it would be good for your LA magazine to have the tests done with large voltage drops across the caps.... as with filters -- better at showing the cap user what is the real world useage distortion.
As the tests have been shown to date, users might get the wrong impression. It wont change the the fact that films are better than polar types and certain films are better than others. But you will have a better knowledge of what the real magnitude of the distortion is in applications other than coupling.
THx-RNMarsh
Last edited:
Road map as usual.
The Fetlington shown as this is more or less all there is to it and worth showing for it's simplicity. I never knew how L FET's are drawn so no idea if right. Take letters as what you should know( N,P ,drain). It will be killer powerful. Even gain will be good. I suspect 90% as opposed to 78% for MOSFET alone. Bipolar EF about 93%.
I would like to hear a working Fetlington done this way in an amp.... fet driving speaker up to some level where bjt comes in.
THx-RNMarsh
All true, but the reality is, nobody does it, and does a decent write-up, so far
Jan
Samuel Groner has done plenty.
I would like to hear a working Fetlington done this way in an amp.... fet driving speaker up to some level where bjt comes in.
THx-RNMarsh
The last circuit would be hard (impossible)? to stably bias for T. If you think about it the reverse (Hafler style) is relatively easy since the DC bias spreader error is across MOSFET gates, relatively low gm. Switched the error is across Vbe's doubling current for each 36mV give or take.
- Status
- Not open for further replies.