That reminds me, I just moved it around yesterday: The gentleman who wrote the very first technical article in the very first JAES issue. He retired in the local area, right here. We was liquidating his gear, and I bought his personally built Dynaco 150 amp from him. Had an afternoon of great talks with him before I left. He was the one who was responsible for the first aircraft ILS systems in Britain, during the war. He used a spitfire ILS as an AM tuner. (think I told that story before)
Keantoken,
One atmosphere is 14.9 lbs of pressure. Vacuum can go as high as about 29" of mercury which is more than one atmosphere. It would be much safer than having to heat any kind of liquid. Get that wrong with a fast temperature change and things start to explode and you don't want glass flying your way. Vacuum impregnation is used all the time.
One atmosphere is 14.9 lbs of pressure. Vacuum can go as high as about 29" of mercury which is more than one atmosphere. It would be much safer than having to heat any kind of liquid. Get that wrong with a fast temperature change and things start to explode and you don't want glass flying your way. Vacuum impregnation is used all the time.
Yet, how many people realize the importance of field effects, and bother to make their designs fully mirror complementary in the physical layout? Due to tracing this can be a serious hassle, yet critical when it comes to dropping out transient distortions. Thus not only worth the bother and headaches to make it physically symmetrical and thought out that way..but ..critical to the attempt, integral to the attempt. Compacting designs can be critical, yes, but this other issue looms larger, IMO.
Last edited:
Something like this ..... ? ....... :I spent some time looking for "period doubling" in op amps just at the brink
of oscillation. No luck.
http://elth.ucv.ro/fisiere/anale/2007/47.pdf
There are patents that have been posted before from 1966, but there is no point this thread has returned to "nothing done in the last 40yr. is new". I think I might have had enough of this.
I don't claim to have been the only inventor of the complementary differential input stage, just one of them. Jon Iverson did it independently of me, and perhaps even before me. Perhaps there were others.
Patents that I have seen earlier have only a superficial resemblance to the TRUE complementary differential input stage and completely impractical for audio.
What I claim is: I developed on my own, the complementary differential bipolar input stage in 1968 in pursuance of a solid state amp to match or beat the finest tube amps. Once I did, I looked around for years for anyone else to develop it. The first time I saw independent development of the complementary differential input stage was with Southwest Technical products, and I commended the designer for that in 1974, when I spoke to him about it. He was the first to publish it in TAS in 1972. By then I had built dozens of amps this way, including a 2000W complementary differential bridge amp, much like what I am describing on this thread at the moment, in 1969 for Ampex Research. The rest I built at Alembic Inc, as a potential replacement for the MAC 75, in 1970-1. Many years later, in the mid 80's did I find that Jon Iverson also used the complementary differential input stage and that he kept it 'confidential' just like I did, for the same reasons. I have great respect for ANYBODY who can think up this sort of topology on their own, without looking at someone else's schematic.
Patents that I have seen earlier have only a superficial resemblance to the TRUE complementary differential input stage and completely impractical for audio.
What I claim is: I developed on my own, the complementary differential bipolar input stage in 1968 in pursuance of a solid state amp to match or beat the finest tube amps. Once I did, I looked around for years for anyone else to develop it. The first time I saw independent development of the complementary differential input stage was with Southwest Technical products, and I commended the designer for that in 1974, when I spoke to him about it. He was the first to publish it in TAS in 1972. By then I had built dozens of amps this way, including a 2000W complementary differential bridge amp, much like what I am describing on this thread at the moment, in 1969 for Ampex Research. The rest I built at Alembic Inc, as a potential replacement for the MAC 75, in 1970-1. Many years later, in the mid 80's did I find that Jon Iverson also used the complementary differential input stage and that he kept it 'confidential' just like I did, for the same reasons. I have great respect for ANYBODY who can think up this sort of topology on their own, without looking at someone else's schematic.
Very useful. Think about a magnetic damper, a permanent magnet next to a spinning copper disc. (or a spinning magnet in an aluminum cup, the old car speedometer trick). At zero velocity, there are no eddy currents within the copper. But the faster the copper spins, the more eddy currents, so the more force applied to the magnet and copper disc. If one were to cut radial slots in the copper to remove the eddy current path, the copper will spin freely.
Now, think of a voice coil with two signals applied, a low frequency and a high frequency. The low frequency is causing the coil to move in the slot. The high frequency signal is seeing the iron or copper ring moving with respect to the coil. IOW, the eddy losses (and forces) of the hf signal will be dependent on the velocity of the coil in the gap. Intermodulation.
Cutting out the eddy forces and losses by slitting as I drew removes this velocity dependent coupling and intermodulation, while compromising reluctance only by the percentage of sectional area lost. EDM using 2 mil wire spaced 30 mils apart drops increases reluctance about 6% in the flux direction, while almost eliminating eddies. Using glued laminations in the "easy" direction will increase reluctance by a factor of 2 or 3, it's a very anisotropic design.
No. The intent of vacuum impregnation is to create a void free embedment or fill. The process involves 3 basic steps. (full vacuum is 0 PSIA, room is 15 PSIA) "A" is absolute.
1. Pull a vacuum on the mixed warmed epoxy (or oil). The purpose of this is to outgas any entrained air or other gases. This must be done at a pressure lower than the final process. If the subsequent process has a lower pressure, it will further outgass as you transfer the liquid. If the de-airing vessel is too tall, there will be a pressure gradient in the liquid, and the top will de-air more than the bottom. For example, if the epoxy has twice the density of water, a one foot column will have 1 PSIA at the bottom even if you have 0 PSIA at the top. If you de-air by subjecting the object to the vacuum after fill, there will still be a vertical pressure gradient, so you may want to put the object on it's side to reduce vertical height. Once the de-airing has slowed significantly, the liquid is ready.
2. The object must be de-aired and dried. Typically, full vacuum at a temp in excess of 100oC (moisture removal). With vacuum maintained, reduce object temp to impregnation temp.
3. Open a valve from the mix chamber to the object. Best to go from bottom of mix to bottom of object. This step requires the object be at or below mix pressure.
4. Pull object to lowest vacuum desired, to complete the de-air.
5. When satisfied, bring object pressure up to cure value, this can easily be room pressure.
This final step is important. It is better to cure the epoxy with higher pressure, as when you increase the pressure, any bubbles left will contract. Some will put the object at high positive pressure at this time to further reduce the bubbles. This will depend heavily on the epoxy, it's pot life, the object dimensions, the interstitial spacing size, and whether the epoxy outgasses volatiles as a consequence of curing. Sy knows more about that stuff, I'm just the mechanic...
For at home, I'd de-air the epoxy warm, fill the object container, pull a vacuum on it while increasing the temp to curing temp, and when it gets to temp, release vacuum. Make sure this is done before gelling.
jn
ps. sorry for the length of the post.
Last edited:
John, I have a design challenge for you and Dick. A <10 ppm triode only 20dB line stage that drives 600 Ohms at 7V rms. Several off the shelf IC's will do this.
I find comments like that simply out of context. With no consideration of available gain, application, impedance levels, load drive, they are just conversation.
This is one topology often favored by Ayre, the Blowtorch, Constellation, and perhaps Nelson Pass. It is more linear than a normal 2 stage design, but it has gain limitations. Perfect for open loop designs. We consider this a 1-1.5 stage design. Almost minimum, only a single triode can do better.
Last edited:
JN,
That would be easier with an autoclave but we are talking about a home here so didn't talk about it. I never saw what material he was trying to use for impregnation or what he was trying to impregnate, just tried to steer him away from his pressure cooker idea. We don't need to make any Boston bombers here!
I'll have to think about your layered lamination model here, it sounds like it would be very costly for any type of production. I have been getting fairly adept at using the FEMM magnetic analysis software, still more to learn but I am doing it. Not sure how exactly to model your idea yet but I think I can figure it out. Would you need to have the entire working area of the pole piece made this way or would a section built that way achieve the blocking of the eddy currents? In other words does it need to cover the entire p to p excursion? I guess a simple picture of the assembly would go a long way to understanding the construction, a die cutter could make the rings but then you would still need to slit those if I am following you correctly? You also said this was a patented design earlier.
That would be easier with an autoclave but we are talking about a home here so didn't talk about it. I never saw what material he was trying to use for impregnation or what he was trying to impregnate, just tried to steer him away from his pressure cooker idea. We don't need to make any Boston bombers here!
I'll have to think about your layered lamination model here, it sounds like it would be very costly for any type of production. I have been getting fairly adept at using the FEMM magnetic analysis software, still more to learn but I am doing it. Not sure how exactly to model your idea yet but I think I can figure it out. Would you need to have the entire working area of the pole piece made this way or would a section built that way achieve the blocking of the eddy currents? In other words does it need to cover the entire p to p excursion? I guess a simple picture of the assembly would go a long way to understanding the construction, a die cutter could make the rings but then you would still need to slit those if I am following you correctly? You also said this was a patented design earlier.
- Status
- Not open for further replies.