The folks I know at JBL is dropping with age and faster turnover. Doug was always a good dinner at trade shows. (I have been told I am down to just being a top 20 dealer. Once at a Crown show they had a new brochure. Looking at it I remarked that all the venues mentioned in the US were my projects. The president jumped up grabed an engineer and confirmed that.)
Certainly no one is using 2" wafers, I don't think the starting material is available anymore (maybe for some very exotic devices). We got a quote today for some specialty JFET's for an experiment $290 @ 1-9 in die form
My last batch of LM3X0 pieces were also marked 7915. They are measuring 20 dB PSSR @ 6k. That is why I was pleased to see similar results in Linear Audio. I expect as devices move to newer processes for the published parameters to improve. I will not show these measurements formally as too many folks don't understand that what today is the "best" part many not be next week and what was not impressive may get much better.
Richard, you really did get spoiled working for LLL. I remember 29 years ago,(1984) calling you up about a high voltage avalanche array that we used to control an optical shutter for a laser. My design info first came from previous work done at LLL or LBL, and when I described what I was doing, you sort of laughed that my approach was very antiquated or primitive, and that I should throw a lot of money at the new and 'better' approach then used.
You see, even though I was part of a design team making a medical laser, I did NOT have an infinite source of funds, and we were not building nuclear bombs, so we did not need more than the array of high voltage transistors to avalanche on trigger.
Yes, we made the 'primitive' controller work, and I saved the company some money.
It is much the same with custom jfets. It is RARE that one needs a custom sort of jfet, unless again, one is making extremely critical designs.
Also, the very idea that somebody will make a CUSTOM jfet for a few hundred dollars is equally absurd. Perhaps a special select of jfets is what is required.
In any case, it might be interesting to discuss how to use jfets more flexibly, so that we can still make pretty good audio circuits with relatively cheap and available jfets.
You see, even though I was part of a design team making a medical laser, I did NOT have an infinite source of funds, and we were not building nuclear bombs, so we did not need more than the array of high voltage transistors to avalanche on trigger.
Yes, we made the 'primitive' controller work, and I saved the company some money.
It is much the same with custom jfets. It is RARE that one needs a custom sort of jfet, unless again, one is making extremely critical designs.
Also, the very idea that somebody will make a CUSTOM jfet for a few hundred dollars is equally absurd. Perhaps a special select of jfets is what is required.
In any case, it might be interesting to discuss how to use jfets more flexibly, so that we can still make pretty good audio circuits with relatively cheap and available jfets.
Well JC, what size company has an R&D budget of $1 Billion a year? yes, spoiled burning thru 25 mill a month ( I didnt do it alone - had 5000 LLNL Scientists/engineers/techs/employess helping). Reaching beyond SOTA. Do that for 25 years and you get pretty far ahead of the game. Where do you think oil shale, retort and coal gasification R&D came from? Lasers and DNA computing and particle accelerators/colliders and all the rest? LLNL, Sandia etc. Hey. Someones got to do the hard risky stuff that results in technology transfers to industry.
Not a lot applies to audio and music, though. So, its been my hobby. I am for learning to use jFEt more flexibly. Whatya got for us? Thx-RNMarsh
Not a lot applies to audio and music, though. So, its been my hobby. I am for learning to use jFEt more flexibly. Whatya got for us? Thx-RNMarsh
Last edited:
I don't disagree with you, Richard. In fact, I had a little of that when I was with Ampex Research in 1969. I still have some planar power complementary transistors that I bought for an amp project, that cost something like $100 ea at the time. Mil quality, gold plated, stud mount, very fast. Those were the days!
The subject that I wanted to talk about is:
Using combinations of Jfets, with other parts, to make 'super parts' that may have even once been made, but it might be cheaper to series and-or parallel devices and get similar results. This concept goes back more than 40 years to the FETRON which was a vacuum tube replacement, composed of a low input capacitance jfet cascoded with a 200-400V jfet (available at the time) for the voltage holdoff. This was originally used to replace real vacuum tubes in VTVM's made by a number of companies.
Today, we might use a MOSFET, instead of the high voltage jfet to get similar results. Heck, today we could make a solid state direct drive for STAX phones with a dual jfet cascoded by a pair of 800V mosfets, and get pretty good results. There are more combinations that we could talk about.
Using combinations of Jfets, with other parts, to make 'super parts' that may have even once been made, but it might be cheaper to series and-or parallel devices and get similar results. This concept goes back more than 40 years to the FETRON which was a vacuum tube replacement, composed of a low input capacitance jfet cascoded with a 200-400V jfet (available at the time) for the voltage holdoff. This was originally used to replace real vacuum tubes in VTVM's made by a number of companies.
Today, we might use a MOSFET, instead of the high voltage jfet to get similar results. Heck, today we could make a solid state direct drive for STAX phones with a dual jfet cascoded by a pair of 800V mosfets, and get pretty good results. There are more combinations that we could talk about.
Almost -- I was a Sr. EE Technical Coordinator on some very large projects. Interfacing with a lot of people from a multitude of disiplines. Fortunately, I didnt have to get votes to keep my job! Just being really good at it was enough.
Richard, I just talked to Jack Bybee about his former project with the government.
He quoted 16 billion dollars for that project, so I asked him why so much? This was the project where we had to know where or quickly locate Russian Atomic missile subs under water before they could launch their missiles. We sort of 'got caught with our pants down', (not the first time) when the Russians one-upped us! Let me congratulate the USSR for doing so, at this much later time, quite a feat!
Anyway, we had to throw money at the problem, just like we did with going to the Moon. That is what made silicon valley so powerful and advanced in those days. You know: MONEY!
He quoted 16 billion dollars for that project, so I asked him why so much? This was the project where we had to know where or quickly locate Russian Atomic missile subs under water before they could launch their missiles. We sort of 'got caught with our pants down', (not the first time) when the Russians one-upped us! Let me congratulate the USSR for doing so, at this much later time, quite a feat!
Anyway, we had to throw money at the problem, just like we did with going to the Moon. That is what made silicon valley so powerful and advanced in those days. You know: MONEY!
The bigger it is, the easier it goes down.
How much it would cost to send back J. B.... on Mars with a quantum powered engine ?
I would prefer we talk about combinations of Jfets, indeed.
Last edited:
$16 billion for a 0.02 ohm resistor? And I thought $600 hammers and toilet seats were bad.
se
Research And Development......
JJBQ4zPvhtIw~~60_1.JPG?set_id=880000500F)
I was given one yesterday....some hammers are priceless.........
Dan.
I was given one yesterday....some hammers are priceless.........
Dan.
Last edited:
yep, I was thinking about that the other day re stax, was actually thinking of a direct digital USB->DSD->Gate driver->LDMOS RF fet STAX amp, but was reminded that most or at least many of these Fets have significant Ciss/Coss and with the ES driver being so low impedance top end distortion is an issue. thats not to say its not doable, but many of the new higher voltage fets are higher capacitance than desirable for such things.
thanks, interesting, nice parts but woah! did you tell them to go jump or are you getting some? if thats what they charge for the wire-bondable parts, what do they charge for the packaged ones?
its getting more and more ridiculous, guess he had to adjust the amount for inflation so it still sounded very impressive.
John, your friend is having a laugh
Add some extra zeros, and Mr Bybee might have been just the right man for the SDI job in the early '80s.
("the Bigger it is", speaking of throwing Gobs of Meuney at General-D, and silicon valley, for a Limp Phal..anx ciws : USS Stark)
I don't know about $16 billion dollar programs in a military research and development but I do understand the $600.00 toiler seat or something like it. Until you see all the documentation and layers of inspection required to produce a part for an aerospace application you don't have any idea how expensive it is to actually produce some of these parts. Specialized tooling using specific materials with exacting quality control do not come cheap. I have had to estimate the cost to produce and bid on open government military projects. What often would catch you off guard was the particular material that was specified to produce a part. Recreating a part that has not been made in years does not mean that you can use a new modern material to produce that part. I have had to get quotations for producing a composite part that would require the use of a particular composite prepreg material. The part may require as little as one or two yards of material but you have to have the original producer of the material make a run of that material to use. Now you are purchasing a minimum of 100 yards of material at a cost of 20 to 30 thousand dollars and up, it is something that may not have been produced for 20 or 30 years and you have to pay to play. There is no old material, it is obsolete or out of date. You can not substitute a new material without going through a complete re-qualification of the part. This could take forever to do and cost more than producing the part with the expensive special order material. There is just no way to lower the cost and meet the requirements. I also was in the position to have to reproduce tooling that was no longer available for production of a part. Having to go all the way back to the original mylar drawings to reproduce a set of tools to make a one off part for an older plane sitting on the ground because somebody drove over an access door with a forklift and the plane cannot fly without the part. I have done that where we called it AOG, aircraft on the ground. Or when I had to reproduce a part for a B52 bomber that hadn't been made in years, Boeing will tell the government to go pound sand, buy a new plane, but someone has to service those planes. I personally built over 200 sets of tooling for KC135 planes for internal and external aluminum bonded panels. That cost somebody lots of money to keep those planes flying around the world.
On a separate track I was talking to a friend the other day and he was telling me about the topology of the old Spectra Sonics amplifiers of the 70's. He was telling me that in that amplifier that the amplifier was built with a unique circuitry. Instead of a differential input pair of transistors the amplifier was built with a differential output. I assume it could have also had the differential input besides the output but don't know. Could somebody comment on that, how that is accomplished and why there aren't more amplifiers with the outputs done this way? I assumed from the conversation that the amplifier was a mono channel card, not a stereo output.
On a separate track I was talking to a friend the other day and he was telling me about the topology of the old Spectra Sonics amplifiers of the 70's. He was telling me that in that amplifier that the amplifier was built with a unique circuitry. Instead of a differential input pair of transistors the amplifier was built with a differential output. I assume it could have also had the differential input besides the output but don't know. Could somebody comment on that, how that is accomplished and why there aren't more amplifiers with the outputs done this way? I assumed from the conversation that the amplifier was a mono channel card, not a stereo output.
It doesn't look like it could go into a product, but it still might be the only thing for a research project. They have one set up with a feedback cap (.03pF) and reset switch for use as a charge amp. To be fair these are pretty special and even require special bonding to accommodate the tiny geometry.
I have some personal knowledge there. The $600 pricetag included amortization of the tooling.
oh I agree I had a good read of the datasheets, I briefly started to convince myself I needed some of the 4 lead high IDSS parts for a cost no object IV stage (I dont have access to wirebonding on parts I could lose inside a teacup)
- Status
- Not open for further replies.