|
HP8640B output stage
Dear all
I have seen posts in the past about this issue My friend has acquired a unit with the output stage defunct. Any ideas about what the problem is likely to be and the solution? Thanks Richard |
Richard Hosking wrote:
Dear all I have seen posts in the past about this issue My friend has acquired a unit with the output stage defunct. Any ideas about what the problem is likely to be and the solution? Thanks Richard Try the test-equipment reflector at qth.net Bob, W6NBI |
Richard Hosking wrote:
Dear all I have seen posts in the past about this issue My friend has acquired a unit with the output stage defunct. Any ideas about what the problem is likely to be and the solution? Thanks Richard Try the test-equipment reflector at qth.net Bob, W6NBI |
In article ,
says... Dear all I have seen posts in the past about this issue My friend has acquired a unit with the output stage defunct. Any ideas about what the problem is likely to be and the solution? Thanks Richard They tend to get fried by reverse power application. Best solution would be to replace it with a part from a donor unit. Next bet thing would probably be a 50-ohm MMIC like a GALI-5 that's capable of similar gain and output power. I don't remember if the 8640B's output-power sampler is built into its output stage, though; if it is, then some careful hacking might be necessary. -- jm ------------------------------------------------------ http://www.qsl.net/ke5fx Note: My E-mail address has been altered to avoid spam ------------------------------------------------------ |
I'm pretty sure it's just an amplifier. I took a very quick look at the schematic this morning for my HP8640M, which should be the same. I believe it's 18dB gain. As you say, it must be capable of the output power. GALI-51 would be a bit closer match for the gain, but not quite enough output power? Cheers, Tom Parallel a couple. W4ZCB |
I'm pretty sure it's just an amplifier. I took a very quick look at the schematic this morning for my HP8640M, which should be the same. I believe it's 18dB gain. As you say, it must be capable of the output power. GALI-51 would be a bit closer match for the gain, but not quite enough output power? Cheers, Tom Parallel a couple. W4ZCB |
Tom Bruhns wrote:
"Ian White, G3SEK" wrote in message ... Is there also an issue about harmonic content? ISTR there is no significant harmonic filtering after that amp, so any replacement needs to be operating well below its quoted maximum power output (which will be at 1dB compression, where significant harmonic output is guaranteed). Depending on the detailed schematic, might the input and output match also be important? Are any filters or attenuators relying on this amp for their wideband 50R source load or source? If there is no fixed attenuator pad at the output, then the variable attenuator is relying on the accuracy of the amp's output impedance (at least at high output levels, ie low attenuation). This might be a case for using push-pull, to reduce the even harmonics, if you can find suitable wideband transformers. Four GALIs of some kind in push-pull parallel might do the job nicely... but if you want to get it as right as HP did in the original design, it ain't trivial. I shouldn't have said that - I'm now too scared to switch on my 8640A! All good points, Ian. Indeed there's no filtering after the amplifier, so if you want to keep the harmonic content low, you will want a decent amplifier there. Also, the output power should be at least +23dBm if you want to have the same performance as the original. However, I believe the output is leveled (or at least monitored), so the output impedance is irrelevant, except perhaps on the highest range where the source impedance may depend on the amplifier output impedance. Good point. I could have a closer look at the schematic (and in the HP Journal article about the 8640B) if anyone has a serious need to know some more details. Are those old HPJs on the web anywhere? It may have been in this group a while ago that mention was made of the way the 8640 generates its output bands: an oscillator from about 250 to 500MHz, with digital divide-by-2 stages followed by LC filters to attenuate the harmonics. Someone said, I believe, that there is only one filter per octave. That is not the case with my 8640M, which on the higher bands has two filters per octave, one for the low half and one for the high half of the band, which are somehow automatically switched. I didn't look into just how the switch point was determined. There's a mechanical switch on the slide-rule tuning mechanism. But the true glory of the 8640 series is the way it keeps the FM deviation constant when more dividers are switched in. Who else but HP would have used a differential gearbox? -- 73 from Ian G3SEK 'In Practice' columnist for RadCom (RSGB) Editor, 'The VHF/UHF DX Book' http://www.ifwtech.co.uk/g3sek |
Tom Bruhns wrote:
"Ian White, G3SEK" wrote in message ... Is there also an issue about harmonic content? ISTR there is no significant harmonic filtering after that amp, so any replacement needs to be operating well below its quoted maximum power output (which will be at 1dB compression, where significant harmonic output is guaranteed). Depending on the detailed schematic, might the input and output match also be important? Are any filters or attenuators relying on this amp for their wideband 50R source load or source? If there is no fixed attenuator pad at the output, then the variable attenuator is relying on the accuracy of the amp's output impedance (at least at high output levels, ie low attenuation). This might be a case for using push-pull, to reduce the even harmonics, if you can find suitable wideband transformers. Four GALIs of some kind in push-pull parallel might do the job nicely... but if you want to get it as right as HP did in the original design, it ain't trivial. I shouldn't have said that - I'm now too scared to switch on my 8640A! All good points, Ian. Indeed there's no filtering after the amplifier, so if you want to keep the harmonic content low, you will want a decent amplifier there. Also, the output power should be at least +23dBm if you want to have the same performance as the original. However, I believe the output is leveled (or at least monitored), so the output impedance is irrelevant, except perhaps on the highest range where the source impedance may depend on the amplifier output impedance. Good point. I could have a closer look at the schematic (and in the HP Journal article about the 8640B) if anyone has a serious need to know some more details. Are those old HPJs on the web anywhere? It may have been in this group a while ago that mention was made of the way the 8640 generates its output bands: an oscillator from about 250 to 500MHz, with digital divide-by-2 stages followed by LC filters to attenuate the harmonics. Someone said, I believe, that there is only one filter per octave. That is not the case with my 8640M, which on the higher bands has two filters per octave, one for the low half and one for the high half of the band, which are somehow automatically switched. I didn't look into just how the switch point was determined. There's a mechanical switch on the slide-rule tuning mechanism. But the true glory of the 8640 series is the way it keeps the FM deviation constant when more dividers are switched in. Who else but HP would have used a differential gearbox? -- 73 from Ian G3SEK 'In Practice' columnist for RadCom (RSGB) Editor, 'The VHF/UHF DX Book' http://www.ifwtech.co.uk/g3sek |
Looking at a cicuit, this issues seem to be
The attenuator presumably relies on the output impedance being 50 ohms The hybrid has a gain of 16dB with an output of about +23dBm. Monitoring and AGC is separate from the hybrid, so all it has to do is provide the gain and output! Supply voltage is 44V There is no post amplifier filtering, so harmonics will depend on the amp. None of the conventional MMICs seem to fit directly - it might be necessary to make up something from discrete components Richard Tom Bruhns wrote in message m... "Ian White, G3SEK" wrote in message ... Is there also an issue about harmonic content? ISTR there is no significant harmonic filtering after that amp, so any replacement needs to be operating well below its quoted maximum power output (which will be at 1dB compression, where significant harmonic output is guaranteed). Depending on the detailed schematic, might the input and output match also be important? Are any filters or attenuators relying on this amp for their wideband 50R source load or source? If there is no fixed attenuator pad at the output, then the variable attenuator is relying on the accuracy of the amp's output impedance (at least at high output levels, ie low attenuation). This might be a case for using push-pull, to reduce the even harmonics, if you can find suitable wideband transformers. Four GALIs of some kind in push-pull parallel might do the job nicely... but if you want to get it as right as HP did in the original design, it ain't trivial. I shouldn't have said that - I'm now too scared to switch on my 8640A! All good points, Ian. Indeed there's no filtering after the amplifier, so if you want to keep the harmonic content low, you will want a decent amplifier there. Also, the output power should be at least +23dBm if you want to have the same performance as the original. However, I believe the output is leveled (or at least monitored), so the output impedance is irrelevant, except perhaps on the highest range where the source impedance may depend on the amplifier output impedance. I could have a closer look at the schematic (and in the HP Journal article about the 8640B) if anyone has a serious need to know some more details. It may have been in this group a while ago that mention was made of the way the 8640 generates its output bands: an oscillator from about 250 to 500MHz, with digital divide-by-2 stages followed by LC filters to attenuate the harmonics. Someone said, I believe, that there is only one filter per octave. That is not the case with my 8640M, which on the higher bands has two filters per octave, one for the low half and one for the high half of the band, which are somehow automatically switched. I didn't look into just how the switch point was determined. Cheers, Tom |
Looking at a cicuit, this issues seem to be
The attenuator presumably relies on the output impedance being 50 ohms The hybrid has a gain of 16dB with an output of about +23dBm. Monitoring and AGC is separate from the hybrid, so all it has to do is provide the gain and output! Supply voltage is 44V There is no post amplifier filtering, so harmonics will depend on the amp. None of the conventional MMICs seem to fit directly - it might be necessary to make up something from discrete components Richard Tom Bruhns wrote in message m... "Ian White, G3SEK" wrote in message ... Is there also an issue about harmonic content? ISTR there is no significant harmonic filtering after that amp, so any replacement needs to be operating well below its quoted maximum power output (which will be at 1dB compression, where significant harmonic output is guaranteed). Depending on the detailed schematic, might the input and output match also be important? Are any filters or attenuators relying on this amp for their wideband 50R source load or source? If there is no fixed attenuator pad at the output, then the variable attenuator is relying on the accuracy of the amp's output impedance (at least at high output levels, ie low attenuation). This might be a case for using push-pull, to reduce the even harmonics, if you can find suitable wideband transformers. Four GALIs of some kind in push-pull parallel might do the job nicely... but if you want to get it as right as HP did in the original design, it ain't trivial. I shouldn't have said that - I'm now too scared to switch on my 8640A! All good points, Ian. Indeed there's no filtering after the amplifier, so if you want to keep the harmonic content low, you will want a decent amplifier there. Also, the output power should be at least +23dBm if you want to have the same performance as the original. However, I believe the output is leveled (or at least monitored), so the output impedance is irrelevant, except perhaps on the highest range where the source impedance may depend on the amplifier output impedance. I could have a closer look at the schematic (and in the HP Journal article about the 8640B) if anyone has a serious need to know some more details. It may have been in this group a while ago that mention was made of the way the 8640 generates its output bands: an oscillator from about 250 to 500MHz, with digital divide-by-2 stages followed by LC filters to attenuate the harmonics. Someone said, I believe, that there is only one filter per octave. That is not the case with my 8640M, which on the higher bands has two filters per octave, one for the low half and one for the high half of the band, which are somehow automatically switched. I didn't look into just how the switch point was determined. Cheers, Tom |
"Ian White, G3SEK" wrote in message ...
Tom Bruhns wrote: "Ian White, G3SEK" wrote in message ... .... I could have a closer look at the schematic (and in the HP Journal article about the 8640B) if anyone has a serious need to know some more details. Are those old HPJs on the web anywhere? For a while they were, then there was only an index, and now I'm unable to find anything. Just yesterday, I boxed up a _complete_ set of HPJs to move to our new site, so if there is a specific article you need, please ask and I'll see what I can do (after the move in a couple weeks). I have the HPJ that covers the 8640 at home, though, and could turn part or all of it into a PDF file. Unfortunately, the PDF would be large because it's a set of images, not text-based, from my scanner. I also have the service manual on the 8640M, which is a little different from either the A or B models. It may have been in this group a while ago that mention was made of the way the 8640 generates its output bands: an oscillator from about 250 to 500MHz, with digital divide-by-2 stages followed by LC filters to attenuate the harmonics. Someone said, I believe, that there is only one filter per octave. That is not the case with my 8640M, which on the higher bands has two filters per octave, one for the low half and one for the high half of the band, which are somehow automatically switched. I didn't look into just how the switch point was determined. There's a mechanical switch on the slide-rule tuning mechanism. Presumably it's a bit different in mine, since there is no slide-rule tuning mechanism. I'll have to have a look. But the true glory of the 8640 series is the way it keeps the FM deviation constant when more dividers are switched in. Who else but HP would have used a differential gearbox? In the 1970's when the 8640 was designed, we did, and people like Collins did, but now we don't. There are very few pots or trimcaps in our circuits these days, too, for the same reason--and there are a lot of processors of various sorts, and lots of DACs and the like. I used to think the 8640 phase noise was pretty good, but it doesn't hold a candle to what we do these days. My fun these days is in finding ways to make meaningful measurements 120dB below full scale with data from a 12-bit or 14-bit ADC. :-) Cheers, Tom |
"Ian White, G3SEK" wrote in message ...
Tom Bruhns wrote: "Ian White, G3SEK" wrote in message ... .... I could have a closer look at the schematic (and in the HP Journal article about the 8640B) if anyone has a serious need to know some more details. Are those old HPJs on the web anywhere? For a while they were, then there was only an index, and now I'm unable to find anything. Just yesterday, I boxed up a _complete_ set of HPJs to move to our new site, so if there is a specific article you need, please ask and I'll see what I can do (after the move in a couple weeks). I have the HPJ that covers the 8640 at home, though, and could turn part or all of it into a PDF file. Unfortunately, the PDF would be large because it's a set of images, not text-based, from my scanner. I also have the service manual on the 8640M, which is a little different from either the A or B models. It may have been in this group a while ago that mention was made of the way the 8640 generates its output bands: an oscillator from about 250 to 500MHz, with digital divide-by-2 stages followed by LC filters to attenuate the harmonics. Someone said, I believe, that there is only one filter per octave. That is not the case with my 8640M, which on the higher bands has two filters per octave, one for the low half and one for the high half of the band, which are somehow automatically switched. I didn't look into just how the switch point was determined. There's a mechanical switch on the slide-rule tuning mechanism. Presumably it's a bit different in mine, since there is no slide-rule tuning mechanism. I'll have to have a look. But the true glory of the 8640 series is the way it keeps the FM deviation constant when more dividers are switched in. Who else but HP would have used a differential gearbox? In the 1970's when the 8640 was designed, we did, and people like Collins did, but now we don't. There are very few pots or trimcaps in our circuits these days, too, for the same reason--and there are a lot of processors of various sorts, and lots of DACs and the like. I used to think the 8640 phase noise was pretty good, but it doesn't hold a candle to what we do these days. My fun these days is in finding ways to make meaningful measurements 120dB below full scale with data from a 12-bit or 14-bit ADC. :-) Cheers, Tom |
Tom Bruhns wrote:
There's a mechanical switch on the slide-rule tuning mechanism. Presumably it's a bit different in mine, since there is no slide-rule tuning mechanism. I'll have to have a look. IIRC, there *is* a switch in the tuning mechanism, and it switches filters right at the midpoint of the band. I can't find my copy of the manual to confirm that. *sigh* -- Mike Andrews Tired old sysadmin since 1964 |
Tom Bruhns wrote:
There's a mechanical switch on the slide-rule tuning mechanism. Presumably it's a bit different in mine, since there is no slide-rule tuning mechanism. I'll have to have a look. IIRC, there *is* a switch in the tuning mechanism, and it switches filters right at the midpoint of the band. I can't find my copy of the manual to confirm that. *sigh* -- Mike Andrews Tired old sysadmin since 1964 |
Tom Bruhns wrote:
Are those old HPJs on the web anywhere? For a while they were, then there was only an index, and now I'm unable to find anything. Just yesterday, I boxed up a _complete_ set of HPJs to move to our new site, so if there is a specific article you need, please ask and I'll see what I can do (after the move in a couple weeks). I have the HPJ that covers the 8640 at home, though, and could turn part or all of it into a PDF file. Unfortunately, the PDF would be large because it's a set of images, not text-based, from my scanner. I also have the service manual on the 8640M, which is a little different from either the A or B models. The 8640 article would only be out of historical interest, to see what was in the designers' minds at the time. I downloaded the 8640B manual from the Army site mentioned here a few days ago, but am still looking for a manual for my 8640A (at the same price :-) Someone said, I believe, that there is only one filter per octave. That is not the case with my 8640M, which on the higher bands has two filters per octave, one for the low half and one for the high half of the band, which are somehow automatically switched. I didn't look into just how the switch point was determined. There's a mechanical switch on the slide-rule tuning mechanism. Presumably it's a bit different in mine, since there is no slide-rule tuning mechanism. Beg your pardon - only the A model has the slide-rule dial, but the filter switch is probably the same. I'll have to have a look. You go do that, Tom - mine's right at the bottom of a stack of HP boxes! But the true glory of the 8640 series is the way it keeps the FM deviation constant when more dividers are switched in. Who else but HP would have used a differential gearbox? In the 1970's when the 8640 was designed, we did, and people like Collins did, but now we don't. There are very few pots or trimcaps in our circuits these days, too, for the same reason--and there are a lot of processors of various sorts, and lots of DACs and the like. I used to think the 8640 phase noise was pretty good, but it doesn't hold a candle to what we do these days. True, but it was better than the first generation of synthesized boxes that followed the 8640, so the less-fashionable 8640 series are still very good value for money for amateurs. With HB crystal filters to attenuate the noise sidebands even further, you can make good spot-frequency measurements up to at least 144MHz (see SM5BSZ's website - I'm currently editing an article by Leif, to appear in DUBUS in the fall). My fun these days is in finding ways to make meaningful measurements 120dB below full scale with data from a 12-bit or 14-bit ADC. :-) Yeah, g'luck with that... -- 73 from Ian G3SEK 'In Practice' columnist for RadCom (RSGB) Editor, 'The VHF/UHF DX Book' http://www.ifwtech.co.uk/g3sek |
Tom Bruhns wrote:
Are those old HPJs on the web anywhere? For a while they were, then there was only an index, and now I'm unable to find anything. Just yesterday, I boxed up a _complete_ set of HPJs to move to our new site, so if there is a specific article you need, please ask and I'll see what I can do (after the move in a couple weeks). I have the HPJ that covers the 8640 at home, though, and could turn part or all of it into a PDF file. Unfortunately, the PDF would be large because it's a set of images, not text-based, from my scanner. I also have the service manual on the 8640M, which is a little different from either the A or B models. The 8640 article would only be out of historical interest, to see what was in the designers' minds at the time. I downloaded the 8640B manual from the Army site mentioned here a few days ago, but am still looking for a manual for my 8640A (at the same price :-) Someone said, I believe, that there is only one filter per octave. That is not the case with my 8640M, which on the higher bands has two filters per octave, one for the low half and one for the high half of the band, which are somehow automatically switched. I didn't look into just how the switch point was determined. There's a mechanical switch on the slide-rule tuning mechanism. Presumably it's a bit different in mine, since there is no slide-rule tuning mechanism. Beg your pardon - only the A model has the slide-rule dial, but the filter switch is probably the same. I'll have to have a look. You go do that, Tom - mine's right at the bottom of a stack of HP boxes! But the true glory of the 8640 series is the way it keeps the FM deviation constant when more dividers are switched in. Who else but HP would have used a differential gearbox? In the 1970's when the 8640 was designed, we did, and people like Collins did, but now we don't. There are very few pots or trimcaps in our circuits these days, too, for the same reason--and there are a lot of processors of various sorts, and lots of DACs and the like. I used to think the 8640 phase noise was pretty good, but it doesn't hold a candle to what we do these days. True, but it was better than the first generation of synthesized boxes that followed the 8640, so the less-fashionable 8640 series are still very good value for money for amateurs. With HB crystal filters to attenuate the noise sidebands even further, you can make good spot-frequency measurements up to at least 144MHz (see SM5BSZ's website - I'm currently editing an article by Leif, to appear in DUBUS in the fall). My fun these days is in finding ways to make meaningful measurements 120dB below full scale with data from a 12-bit or 14-bit ADC. :-) Yeah, g'luck with that... -- 73 from Ian G3SEK 'In Practice' columnist for RadCom (RSGB) Editor, 'The VHF/UHF DX Book' http://www.ifwtech.co.uk/g3sek |
"Ian White, G3SEK" wrote in message ...
Tom Bruhns wrote: ....(someone wrote: There's a mechanical switch on the slide-rule tuning mechanism.) Presumably it's a bit different in mine, since there is no slide-rule tuning mechanism. Beg your pardon - only the A model has the slide-rule dial, but the filter switch is probably the same. Well, unless you consider a Schmitt trigger to be a mechanical switch, mine's not the same... .... My fun these days is in finding ways to make meaningful measurements 120dB below full scale with data from a 12-bit or 14-bit ADC. :-) Yeah, g'luck with that... Works well if the ADC is linear enough and you have enough processing gain. Maybe some day not too far down the road we'll have 20 bit and 24 bit 100Ms/s ADCs, but they're a bit rare right now. Cheers, Tom |
"Ian White, G3SEK" wrote in message ...
Tom Bruhns wrote: ....(someone wrote: There's a mechanical switch on the slide-rule tuning mechanism.) Presumably it's a bit different in mine, since there is no slide-rule tuning mechanism. Beg your pardon - only the A model has the slide-rule dial, but the filter switch is probably the same. Well, unless you consider a Schmitt trigger to be a mechanical switch, mine's not the same... .... My fun these days is in finding ways to make meaningful measurements 120dB below full scale with data from a 12-bit or 14-bit ADC. :-) Yeah, g'luck with that... Works well if the ADC is linear enough and you have enough processing gain. Maybe some day not too far down the road we'll have 20 bit and 24 bit 100Ms/s ADCs, but they're a bit rare right now. Cheers, Tom |
| All times are GMT +1. The time now is 09:48 AM. |
Powered by vBulletin® Copyright ©2000 - 2024, Jelsoft Enterprises Ltd.
RadioBanter.com