|
I was going through some old HP Journals and one of them showed a cross section of the cavity oscillator of the 8640 and it showed HP-21 , usually most HP devices I am aware of either started with HXTR or had a Part number (5086 , 1902
) which is consistent with the part number floated in this thread , so I thought that either this was a special part or that since the drawing was published at the time the 8640 was released that maybe it was the part number for a die level part that was mounted
in a TO18 can and that at the time the drawing was made no official part number had been assigned yet
?
toggle quoted message
Show quoted text
From: [email protected] < [email protected]>
On Behalf Of Tom Lee via groups.io
Sent: Wednesday, April 13, 2022 11:49 PM
To: [email protected]
Subject: Re: [HP-Agilent-Keysight-equipment] HP8640B RF Fails Several Seconds After Power Up
?
The HP-21 may very well have been specially made for the 8640. The only Google hits point to that sig gen. I was once told by an HP oldtimer that the 358xx series was derived from this part (or maybe it was the other way around), but he
was not at all certain. The 35824E is in a TO-72 and has specs that look a lot like the ones I found in that one paper's footnote, but I do not know if the case is tied to the base.
--Tom
--
Prof. Thomas H. Lee
Allen Ctr., Rm. 205
350 Jane Stanford Way
Stanford University
Stanford, CA 94305-4070
On 4/13/2022 21:08, Lothar baier wrote:
I am somewhat curious about the part numbering , HP-21 doesn’t fit into any of the numbering conventions for a part or product so I was wondering if this potentially was a part that was especially designed for the 8640 series .
?
?
?
Yes, it would be fun to crank out a run of these transistors. One wafer would yield enough to keep our 8640s working pretty much for as long as anyone would want. I've always fantasized about a desktop fab that could do something like that.
Surprising how difficult it is to replicate 1970s technology with 21st-century tools!
-- Cheers,
Tom
--
Prof. Thomas H. Lee
Allen Ctr., Rm. 205
350 Jane Stanford Way
Stanford University
Stanford, CA 94305-4070
On 4/13/2022 19:47, Lothar baier wrote:
Apologies I didn’t see this one as for some reason the message didn’t show in my email box but after going through the thread on the group I found it !
Most SiGe devices I know of are designed for lower voltages probably because nowadays with everything being battery operated higher voltage devices are no longer a thing but that’s just a theory !
Unfortunately I am no longer at a job dealing with MMIC design but one could conceivably find a silicon or SiGe Process that accommodates those specs and jump on a pizza mask to get some dies and then bond them in a TO-18 can ……….? , with
a ton of 8640 out there it might be worth considering
?
?
I gave the specs earlier in the thread. Knowing those in advance of recommending a device improves SNR.
Tom
--
Prof. Thomas H. Lee
Allen Ctr., Rm. 205
350 Jane Stanford Way
Stanford University
Stanford, CA 94305-4070
On 4/13/2022 18:36, Lothar baier wrote:
Since I don’t know the specs of the original device I cant really judge which device would be feasible , I chose the BFP740 because I used it in the past in low phase-noise oscillator circuits , there is a slew of other devices out there
with lower FT
?
A BFP740 is, in a sense, much too fast. By that I mean that other parameters have been traded off in exchange for high ft. BVCBO is too low, and so is the power dissipation spec, for example.
--Tom
--
Prof. Thomas H. Lee
Allen Ctr., Rm. 205
350 Jane Stanford Way
Stanford University
Stanford, CA 94305-4070
On 4/13/2022 18:16, Lothar baier wrote:
For best phase-noise I would recommend finding a SiGe Device like the BFP740 or similar , Infineon and NXP both make great devices
?
?
Re: Ft overhead: Keep in mind that you need more gain than the minimum that sustains oscillation in the absence of a load -- you need enough to support the power delivered to a load. On top of that, you don't want the frequency of oscillation
to be affected too much by the transistor's own phase shift -- you want the resonator to control it. All those considerations argue for a healthy ft margin. I'm sure you could sub a somewhat slower transistor and still get output, but you will sacrifice some
of the famous (and hard-fought) stability for which the HP8640B is prized.
-- Cheers
Tom
--
Prof. Thomas H. Lee
Allen Ctr., Rm. 205
350 Jane Stanford Way
Stanford University
Stanford, CA 94305-4070
On 4/13/2022 16:44, Flannel Tuba wrote:
Thank you, Arie for finding a link to the actual scanned? article! Having the pictures along with he text is quite a nice addition.
Meanwhile, I have ordered a smattering of VHF and UHF transistors from Mouser with marginally similar ratings to those Tom provided. I do wonder about the fT of 5GHz though. The oscillator's range is from 230-550MHz, so I wonder why the extreme frequency overhead
margin. I went ahead and ordered a dozen or so potential replacements having at least the Vcbo of 30v, Vceo of 20v, Vebo of 4v and fT of over 600MHz with several in the 1-5GHz range. A couple are TO92 packages, which I envision just folding the base lead across
the top to make contact with the grounding hex cap nut/cover, but most are SOT-23, which I'll have to come up with an adapter of some sort for. Any ideas on this are more than welcome.?
Well, if I'm lucky I'll have some time in the next few weekends to experiment with possible replacements for the rare and venerable HP 5086-7082.
I'll let you know what I find.
-Scott?
?
?
?
?
?
|
Hmmm , this would indicate to me that the HP-21 indeed applied to a die rather than a packaged part , YTO always used die not packaged devices
?
toggle quoted message
Show quoted text
From: [email protected] < [email protected]>
On Behalf Of Tom Lee via groups.io
Sent: Wednesday, April 13, 2022 11:55 PM
To: [email protected]
Subject: Re: [HP-Agilent-Keysight-equipment] HP8640B RF Fails Several Seconds After Power Up
?
Oh, silly me: I forgot the very paper I've cited, which describes a YTO using that same transistor. So, at least two products used that transistor.
--
Prof. Thomas H. Lee
Allen Ctr., Rm. 205
350 Jane Stanford Way
Stanford University
Stanford, CA 94305-4070
On 4/13/2022 21:48, Tom Lee wrote:
The HP-21 may very well have been specially made for the 8640. The only Google hits point to that sig gen. I was once told by an HP oldtimer that the 358xx series was derived from this part (or maybe it was the other way around), but he
was not at all certain. The 35824E is in a TO-72 and has specs that look a lot like the ones I found in that one paper's footnote, but I do not know if the case is tied to the base.
--Tom
--
Prof. Thomas H. Lee
Allen Ctr., Rm. 205
350 Jane Stanford Way
Stanford University
Stanford, CA 94305-4070
On 4/13/2022 21:08, Lothar baier wrote:
I am somewhat curious about the part numbering , HP-21 doesn’t fit into any of the numbering conventions for a part or product so I was wondering if this potentially was a part that was especially designed for the 8640 series .
?
?
?
Yes, it would be fun to crank out a run of these transistors. One wafer would yield enough to keep our 8640s working pretty much for as long as anyone would want. I've always fantasized about a desktop fab that could do something like that.
Surprising how difficult it is to replicate 1970s technology with 21st-century tools!
-- Cheers,
Tom
--
Prof. Thomas H. Lee
Allen Ctr., Rm. 205
350 Jane Stanford Way
Stanford University
Stanford, CA 94305-4070
On 4/13/2022 19:47, Lothar baier wrote:
Apologies I didn’t see this one as for some reason the message didn’t show in my email box but after going through the thread on the group I found it !
Most SiGe devices I know of are designed for lower voltages probably because nowadays with everything being battery operated higher voltage devices are no longer a thing but that’s just a theory !
Unfortunately I am no longer at a job dealing with MMIC design but one could conceivably find a silicon or SiGe Process that accommodates those specs and jump on a pizza mask to get some dies and then bond them in a TO-18 can ……….? , with
a ton of 8640 out there it might be worth considering
?
?
I gave the specs earlier in the thread. Knowing those in advance of recommending a device improves SNR.
Tom
--
Prof. Thomas H. Lee
Allen Ctr., Rm. 205
350 Jane Stanford Way
Stanford University
Stanford, CA 94305-4070
On 4/13/2022 18:36, Lothar baier wrote:
Since I don’t know the specs of the original device I cant really judge which device would be feasible , I chose the BFP740 because I used it in the past in low phase-noise oscillator circuits , there is a slew of other devices out there
with lower FT
?
A BFP740 is, in a sense, much too fast. By that I mean that other parameters have been traded off in exchange for high ft. BVCBO is too low, and so is the power dissipation spec, for example.
--Tom
--
Prof. Thomas H. Lee
Allen Ctr., Rm. 205
350 Jane Stanford Way
Stanford University
Stanford, CA 94305-4070
On 4/13/2022 18:16, Lothar baier wrote:
For best phase-noise I would recommend finding a SiGe Device like the BFP740 or similar , Infineon and NXP both make great devices
?
?
Re: Ft overhead: Keep in mind that you need more gain than the minimum that sustains oscillation in the absence of a load -- you need enough to support the power delivered to a load. On top of that, you don't want the frequency of oscillation
to be affected too much by the transistor's own phase shift -- you want the resonator to control it. All those considerations argue for a healthy ft margin. I'm sure you could sub a somewhat slower transistor and still get output, but you will sacrifice some
of the famous (and hard-fought) stability for which the HP8640B is prized.
-- Cheers
Tom
--
Prof. Thomas H. Lee
Allen Ctr., Rm. 205
350 Jane Stanford Way
Stanford University
Stanford, CA 94305-4070
On 4/13/2022 16:44, Flannel Tuba wrote:
Thank you, Arie for finding a link to the actual scanned? article! Having the pictures along with he text is quite a nice addition.
Meanwhile, I have ordered a smattering of VHF and UHF transistors from Mouser with marginally similar ratings to those Tom provided. I do wonder about the fT of 5GHz though. The oscillator's range is from 230-550MHz, so I wonder why the extreme frequency overhead
margin. I went ahead and ordered a dozen or so potential replacements having at least the Vcbo of 30v, Vceo of 20v, Vebo of 4v and fT of over 600MHz with several in the 1-5GHz range. A couple are TO92 packages, which I envision just folding the base lead across
the top to make contact with the grounding hex cap nut/cover, but most are SOT-23, which I'll have to come up with an adapter of some sort for. Any ideas on this are more than welcome.?
Well, if I'm lucky I'll have some time in the next few weekends to experiment with possible replacements for the rare and venerable HP 5086-7082.
I'll let you know what I find.
-Scott?
?
?
?
?
?
?
|
Looking through old files i found a company that manufactures TO18 packages and one of the variants they offer has the center pin (base) connected to the case opposed to the standard emitter connected to case configuration!?
Might be possible to either find a die level part and bond it in or fiddle a SOT package part in there?
toggle quoted message
Show quoted text
On Apr 14, 2022, at 00:10, Lothar baier via groups.io <Lothar@...> wrote:
?
Hmmm , this would indicate to me that the HP-21 indeed applied to a die rather than a packaged part , YTO always used die not packaged devices
?
?
Oh, silly me: I forgot the very paper I've cited, which describes a YTO using that same transistor. So, at least two products used that transistor.
--
Prof. Thomas H. Lee
Allen Ctr., Rm. 205
350 Jane Stanford Way
Stanford University
Stanford, CA 94305-4070
On 4/13/2022 21:48, Tom Lee wrote:
The HP-21 may very well have been specially made for the 8640. The only Google hits point to that sig gen. I was once told by an HP oldtimer that the 358xx series was derived from this part (or maybe it was the other way around), but he
was not at all certain. The 35824E is in a TO-72 and has specs that look a lot like the ones I found in that one paper's footnote, but I do not know if the case is tied to the base.
--Tom
--
Prof. Thomas H. Lee
Allen Ctr., Rm. 205
350 Jane Stanford Way
Stanford University
Stanford, CA 94305-4070
On 4/13/2022 21:08, Lothar baier wrote:
I am somewhat curious about the part numbering , HP-21 doesn’t fit into any of the numbering conventions for a part or product so I was wondering if this potentially was a part that was especially designed for the 8640 series .
?
?
?
Yes, it would be fun to crank out a run of these transistors. One wafer would yield enough to keep our 8640s working pretty much for as long as anyone would want. I've always fantasized about a desktop fab that could do something like that.
Surprising how difficult it is to replicate 1970s technology with 21st-century tools!
-- Cheers,
Tom
--
Prof. Thomas H. Lee
Allen Ctr., Rm. 205
350 Jane Stanford Way
Stanford University
Stanford, CA 94305-4070
On 4/13/2022 19:47, Lothar baier wrote:
Apologies I didn’t see this one as for some reason the message didn’t show in my email box but after going through the thread on the group I found it !
Most SiGe devices I know of are designed for lower voltages probably because nowadays with everything being battery operated higher voltage devices are no longer a thing but that’s just a theory !
Unfortunately I am no longer at a job dealing with MMIC design but one could conceivably find a silicon or SiGe Process that accommodates those specs and jump on a pizza mask to get some dies and then bond them in a TO-18 can ……….? , with
a ton of 8640 out there it might be worth considering
?
?
I gave the specs earlier in the thread. Knowing those in advance of recommending a device improves SNR.
Tom
--
Prof. Thomas H. Lee
Allen Ctr., Rm. 205
350 Jane Stanford Way
Stanford University
Stanford, CA 94305-4070
On 4/13/2022 18:36, Lothar baier wrote:
Since I don’t know the specs of the original device I cant really judge which device would be feasible , I chose the BFP740 because I used it in the past in low phase-noise oscillator circuits , there is a slew of other devices out there
with lower FT
?
A BFP740 is, in a sense, much too fast. By that I mean that other parameters have been traded off in exchange for high ft. BVCBO is too low, and so is the power dissipation spec, for example.
--Tom
--
Prof. Thomas H. Lee
Allen Ctr., Rm. 205
350 Jane Stanford Way
Stanford University
Stanford, CA 94305-4070
On 4/13/2022 18:16, Lothar baier wrote:
For best phase-noise I would recommend finding a SiGe Device like the BFP740 or similar , Infineon and NXP both make great devices
?
?
Re: Ft overhead: Keep in mind that you need more gain than the minimum that sustains oscillation in the absence of a load -- you need enough to support the power delivered to a load. On top of that, you don't want the frequency of oscillation
to be affected too much by the transistor's own phase shift -- you want the resonator to control it. All those considerations argue for a healthy ft margin. I'm sure you could sub a somewhat slower transistor and still get output, but you will sacrifice some
of the famous (and hard-fought) stability for which the HP8640B is prized.
-- Cheers
Tom
--
Prof. Thomas H. Lee
Allen Ctr., Rm. 205
350 Jane Stanford Way
Stanford University
Stanford, CA 94305-4070
On 4/13/2022 16:44, Flannel Tuba wrote:
Thank you, Arie for finding a link to the actual scanned? article! Having the pictures along with he text is quite a nice addition.
Meanwhile, I have ordered a smattering of VHF and UHF transistors from Mouser with marginally similar ratings to those Tom provided. I do wonder about the fT of 5GHz though. The oscillator's range is from 230-550MHz, so I wonder why the extreme frequency overhead
margin. I went ahead and ordered a dozen or so potential replacements having at least the Vcbo of 30v, Vceo of 20v, Vebo of 4v and fT of over 600MHz with several in the 1-5GHz range. A couple are TO92 packages, which I envision just folding the base lead across
the top to make contact with the grounding hex cap nut/cover, but most are SOT-23, which I'll have to come up with an adapter of some sort for. Any ideas on this are more than welcome.?
Well, if I'm lucky I'll have some time in the next few weekends to experiment with possible replacements for the rare and venerable HP 5086-7082.
I'll let you know what I find.
-Scott?
?
?
?
?
?
?
|
