Hi Demian,
Somehow I trust the instrument HP made. B&K doesn't generate confidence for me as the 530 isn't what I would call accurate. I used to sell that product line.
I have never seen that HP instrument. I am intrigued with it.
Hi rsavas,
Somehow I trust the instrument HP made. B&K doesn't generate confidence for me as the 530 isn't what I would call accurate. I used to sell that product line.
I have never seen that HP instrument. I am intrigued with it.
Hi rsavas,
How did that happen?
Tek current probe- Easy to understand a missing probe if you see the prices for those. Even used they are really expensive.
On the 530- Absolute accuracy of Ft is a mushy thing but the numbers I get are pretty close to the datasheet values. I discovered that in an EE Times article many years ago. The HP fixture for a network analyzer is quite premium. I believe they had some other instrument, probably part of a production semi test system that could do the same thing.
On the 530- Absolute accuracy of Ft is a mushy thing but the numbers I get are pretty close to the datasheet values. I discovered that in an EE Times article many years ago. The HP fixture for a network analyzer is quite premium. I believe they had some other instrument, probably part of a production semi test system that could do the same thing.
It fell off the back of the truck, that and a few other Tek goodies. One of my more fortunate encounters. When we meet at a BBQ again, I can give you the real story ... and the deal that was made.
Demian's got it right, I can not justify to get the probe since they are expensive, no deals to be had, like I got on the much cheaper passive current probe, P6021. I have used the P6021 for electronic lighting ballast work in the past to measure lamp current.
Rick
Last edited:
Hi Rick,
I really hope to meet you at the next BBQ. We had a family meet in Ottawa for the last one or I would have been there. Next time I'll bring the tube amp I designed. It sounds much better than the Eico HF-87 its based on and that I brought to a previous meet. It's just a prototype right now. Aluminium chassis, bare.
Should be an interesting story.
-Chris
I really hope to meet you at the next BBQ. We had a family meet in Ottawa for the last one or I would have been there. Next time I'll bring the tube amp I designed. It sounds much better than the Eico HF-87 its based on and that I brought to a previous meet. It's just a prototype right now. Aluminium chassis, bare.
Should be an interesting story.
-Chris
Chris, can't wait to see your creation and get my story out.
If you need to borrow a fancy Tek active probe, I can oblige, since I have a few of them sitting on the shelf, they use the old Tek interface, I also have a newer scope with the new I/F but I need the expensive adapter to use those probes with it, more stuff I can't justify the expense. I have a TDS3054B and a DPO4034.
BTW, if you ever work on a Pioneer SX-1980, it probably will need a new PS board as this design is one of those Japanese cookers that you despise, as you know, they mounted it on the bottom with little or no air flow which fries them. I beefed up the design to include bigger HS's to at least make them run a bit cooler and used those 50W MJE1503x bjts.
Cheers
Rick
If you need to borrow a fancy Tek active probe, I can oblige, since I have a few of them sitting on the shelf, they use the old Tek interface, I also have a newer scope with the new I/F but I need the expensive adapter to use those probes with it, more stuff I can't justify the expense. I have a TDS3054B and a DPO4034.
BTW, if you ever work on a Pioneer SX-1980, it probably will need a new PS board as this design is one of those Japanese cookers that you despise, as you know, they mounted it on the bottom with little or no air flow which fries them. I beefed up the design to include bigger HS's to at least make them run a bit cooler and used those 50W MJE1503x bjts.
Cheers
Rick
Last edited:
Hi Rick,
Well, I need to do something about current probes in the future. I'm currently using an HP 428 (not in pictures) for DC current measurements. Big probe, but has a 1 mA FS range.
I have worked on more than a couple. I mostly dislike them for what they do to their owners and I hate making a repair that isn't permanent. I'm sure you can understand. The biggest problem is that the traces can come loose over time with that heat.
Have you thought of drilling some holes in the chassis above the hotter power supply parts? I'm going to bet you that it wouldn't take much to drop the temperature considerably. I've been waiting for another to come in for service so I can try that out. If that works, I'll look forward to doing those units for the simple fact that the repair will last.
-Chris
Well, I need to do something about current probes in the future. I'm currently using an HP 428 (not in pictures) for DC current measurements. Big probe, but has a 1 mA FS range.
I have worked on more than a couple. I mostly dislike them for what they do to their owners and I hate making a repair that isn't permanent. I'm sure you can understand. The biggest problem is that the traces can come loose over time with that heat.
Have you thought of drilling some holes in the chassis above the hotter power supply parts? I'm going to bet you that it wouldn't take much to drop the temperature considerably. I've been waiting for another to come in for service so I can try that out. If that works, I'll look forward to doing those units for the simple fact that the repair will last.
-Chris
Thank you for the flowers 🙂 I do design work as a free lancer: digital, mixed signal, analog. Mainly precision timing, phase noise related stuff etc. May even include some software in homeopathic doses.
Left rack: My phase noise measurement machine in statu nascendi; has lots of delay lines made from many 1:6 coax relays and semi rigid cables, mixers, some ref oscs; HP5370 time interval counter (an array of bad contacts), adjustable digital delay, 8662 synth. generator, local frequency/timing ref, homebrew.
Top: HP quad system power supply, W&G SNA33 spectrum analyzer 30 GHz @1 Hz BW, 53310A modulation domain analyzer(see jitter etc on carriers), homemade isolation transformer 650 VA, normally for scope & fun generator, mW meter, GPS RX 58540A and home brew distributor, Racal counter,TEK current probe & amp 50 MHz middle: 8970B noise figure meter, Anritsu 3GHz pulse generator, 89441A signal analyzer/FFT, Fun-Gen 3325B, 54846B scope 2.3GHz/8GSPS, SMPD signal gen.
bottom: Iwatsu 200 MHz analog scope behind the screen, seldom used since the 54846B, Prom burner, some R&S NGT20 & NGK35 power supplies -like!!- DG8SAQ VNWA, ZVB8 vector network analyzer, 54750 digitizing scope (18 GHz TDR, 20/50 GHz BW), 3478A multimeter, URV35 & Z51 thermal power meter head.
right rack: Lucent GPS, somewhat modded, power supply, SR620 time interval ctr, SMIQ4 generator, 16702B logic analyzer, also has scope cards, but they are clumsy to use. Ayoue 852 chinese hot air, Metcal 5200 solder station -LOVE!!!-
Took quite some time to collect the stuff. It would have been tremendous help if I had that earlier.
Under the microscope is the LTC2500-32 ADC & Beaglebone Black I wrote about in a different thread.
Cheers, Gerhard
Hi Gerhard,
The closest thing I have for measuring jitter is an HP 5372A with 50R pods, I have also a 1M pod with probe in a kit. My Agilent MSO is a 54642D. The Philips PM3365A is there for looking at eye patterns and some analog stuff, I greatly prefer the Agilent. I wish I could afford a current product, 4000 series would be ideal.
You have quite a capable bench there. I can see why it took you years to gather the equipment up.
-Chris
The closest thing I have for measuring jitter is an HP 5372A with 50R pods, I have also a 1M pod with probe in a kit. My Agilent MSO is a 54642D. The Philips PM3365A is there for looking at eye patterns and some analog stuff, I greatly prefer the Agilent. I wish I could afford a current product, 4000 series would be ideal.
You have quite a capable bench there. I can see why it took you years to gather the equipment up.
-Chris
Really nice setup Gerhard, cheers to you, just think what all that stuff costs new 🙂
Chris,
If you want to borrow the TCPA300, no problem, just that we need a probe for it 🙂
IMO, the SX-1980 needs a bunch of holes drilled in the back plate to allow for the heat to escape, allow for air to flow instead of getting trapped inside. One problem is the customers do not want the unit to be altered physically since it can decrease the value.
Rick
Chris,
If you want to borrow the TCPA300, no problem, just that we need a probe for it 🙂
IMO, the SX-1980 needs a bunch of holes drilled in the back plate to allow for the heat to escape, allow for air to flow instead of getting trapped inside. One problem is the customers do not want the unit to be altered physically since it can decrease the value.
Rick
Hi Rick,
-Chris
And there's the rub. Next time I'm in one I'll pay attention to that area and maybe figure something out. This is an extreme problem. Maybe a small fan could be placed inside to circulate the air out existing ventilation paths?
-Chris
I have seen a fan used, mounted on a side but some balk at that too, not OEM. Mounting a fan is a mech problem, power from? AC, as you really do not want any more load on the reg supplies. So either AC and the noise it may inject or load the unregs some more. Problem is there are no/few ventilation path(s), aka our Japanese cooker.
I found a Tek 134 amplifier, probe and 50R terminator on Ebay several years back -- all were ~$120 or in that vicinity. The VNA I have incorporates a calibration routine so the device is adequate for my needs.
Hi Jack,
Now there is luck! If I found a deal like that I'd be all over it.
Hi Rick,
You're probably right. This is a case of anything being better than nothing. I'll have a close look at one of these the next time I see one.
-Chris
Now there is luck! If I found a deal like that I'd be all over it.
Hi Rick,
You're probably right. This is a case of anything being better than nothing. I'll have a close look at one of these the next time I see one.
-Chris
We used to have Agilent for working on eye diagrams, and one TEK CS?8000 for doing the pics for the data sheets; they were simply prettier.
No one in our league can afford the leading edge current products, and also the semi manufacturers think more than twice before they buy. Keysight has now a scope that can do 110 GHz real time. When we built 10 GB/s XFP fiber optic transceivers, we lost more than a month with < 1 bit error per measurement day; we had tho announce measurements to the customers and report the results; They would not accept better results unless they could be properly explained by design changes.
It turned out that was caused by the size of the coupling capacitors; luckily it was the head of development who had decided that 10 nF were enough for 10 GB/s. But the polynomial used for testing had a zero runlength that just was a little bit too long. 100nF 0402-> no problem. If we had anything that could show the bad bit cell, even with < 5 GHz BW but real time, that would have saved a pile of money.
I have also worked for a manufacturer of wafer testers; I had a mixed signal wafer tester all for me alone; maybe once every other week someone would want it for checking sth. The semi makers would not buy one of these without sleeping over it a few nights. Mine had also a AP27-something, type IIRC like an early eprom. It was a bitch to integrate btw; we had to float it.
I fully understand the need to sell crippled scopes; some people are proud when they crack their Rigol and get more bandwidth than they bought; but nowadays the value is in the software and not in the clock rate of some cheap ADC.
In our engineering hall, there were abt. 300 engineers; the hardware group was maybe a good dozen of people. Ridiculous. The number. Not the people. The operating software of the tester was 10 million lines. Its support to use several programming languages on the same job is what now holds Open Office together. They paid millions to SUN to make that happen.
I'm getting too verbose. Have driven 700 Kms today in deluge-like rain. Bed time.
Cheers, Gerhard Excellent Rioja.
Hi Gerhard,
A higher end Agilent scope would have handled eye patterns better. Mine, 54642D, was worth close to $20K US when it was new. That wasn't that long ago, but you would have needed a better DSO to show an eye pattern with the clarity an analogue 'scope can (hence the Philips PM3365A on my bench). I found that the Philips scopes had a far sharper trace than Tektronix products often have. It totally kills my 2465B. I'll change some capacitors and hope that fixes the display issues. If not, I'll sell the darned thing.
-Chris
A higher end Agilent scope would have handled eye patterns better. Mine, 54642D, was worth close to $20K US when it was new. That wasn't that long ago, but you would have needed a better DSO to show an eye pattern with the clarity an analogue 'scope can (hence the Philips PM3365A on my bench). I found that the Philips scopes had a far sharper trace than Tektronix products often have. It totally kills my 2465B. I'll change some capacitors and hope that fixes the display issues. If not, I'll sell the darned thing.
-Chris
I have been dealing with eyepatterns on HDMI for 10+ years. I'm convinced its a racket to sell new scopes. We have a DSA8200 with two dual TDR 20 GHZ plugins. Tek tells us the the 8200 is obsolete and not really supported. Got to get a DSA8300 mainframe to use the current software, and shell out $$$$ for the software to make it useful. And then it turns out the software was not written by Tek. The funny thing is that the TDR I have for the 7854 frame has the same bandwidth. I suspect there is a way to get an eyepattern on the vintage hardware.
Now they tell us to dump the 20 GHz samplers and switch to a 4 port network analyzers and simulate eyepatterns for cables. What HP wants for the real time scope for looking at eyepatterns is shocking. And it will be yesterdays news in a year or so. I could buy a small house for that much $$$ and the house will be worth more in 20 years, not scrap value. Ultimately bit error rate was much more meaningful but those are also expensive.
If there weren't a lot of water between us I may have had some useful 12 Gbps bert stuff to pass along. I think it was scrapped.
Those are very expensive games to play and the competition for consumer (HDMI) or commercial (telecom) high data rate stuff is brutal. Either you win the deals or its millions down the drain.
Now they tell us to dump the 20 GHz samplers and switch to a 4 port network analyzers and simulate eyepatterns for cables. What HP wants for the real time scope for looking at eyepatterns is shocking. And it will be yesterdays news in a year or so. I could buy a small house for that much $$$ and the house will be worth more in 20 years, not scrap value. Ultimately bit error rate was much more meaningful but those are also expensive.
If there weren't a lot of water between us I may have had some useful 12 Gbps bert stuff to pass along. I think it was scrapped.
Those are very expensive games to play and the competition for consumer (HDMI) or commercial (telecom) high data rate stuff is brutal. Either you win the deals or its millions down the drain.
Hi Demian,
If you`re talking to me, nothing but dry land between us. For the low frequency CD eye patterns, I think the 300 series do a pretty good job showing an eye pattern. When I had both the Keysight and Tektronix scopes here for a week, the Keysight won hands down. THey were both easy to set up, and very easy to configure for my lan. The Keysight just handled better and more like an old analogue scope. The Tektronix was noisier and threw off more heat than the Keysight did.
If I won a lottery (and my wife allowed it) the Keysight would be my choice at this point in time.
When you say scrapped, it's such a very sad thing to hear. Hearing about equipment discarded into the trash isn't a fun thing to contemplate for me. It's almost a crime.
-Chris
If you`re talking to me, nothing but dry land between us. For the low frequency CD eye patterns, I think the 300 series do a pretty good job showing an eye pattern. When I had both the Keysight and Tektronix scopes here for a week, the Keysight won hands down. THey were both easy to set up, and very easy to configure for my lan. The Keysight just handled better and more like an old analogue scope. The Tektronix was noisier and threw off more heat than the Keysight did.
If I won a lottery (and my wife allowed it) the Keysight would be my choice at this point in time.
When you say scrapped, it's such a very sad thing to hear. Hearing about equipment discarded into the trash isn't a fun thing to contemplate for me. It's almost a crime.
-Chris
At least my 50 GHz and the TDR plugin are supported by the newer 86xxx communication analyzer series. The 20 GHz are not, at least not officially. But changing the fast units and the TDRs was probably just too much ado for the few units they could sell in addition. Mechanically/electrically they are the same between the series.
So the idea was to buy an El Cheapo HP54750 and go with it as long as it is practical. When I happen to find a nice 86100C/D mainframe I could switch and keep at least the plug ins. Apart of the trigger, there is not much interesting in the mainframe. It's just a windows computer. The 54750 has somehow the smell of UCSD-Pascal. More a belly feeling, I cannot prove that.
wrt eye diagrams: they can tell you a lot of what is wrong with your laser driver. Excessive bias for example leads to an overshoot at the beginning of the bit cell. The Bias pumps a lot of electrons in the laser diode to an elevated state, and when lasing really starts, they all go off immediately. After that event, the optical level collapses to that what is re-pumped on the fly. You can never see that in simulation unless you include the laser diode physics. (Yes, that's simplified. The diode is never completely off.)
\Gerhard
So the idea was to buy an El Cheapo HP54750 and go with it as long as it is practical. When I happen to find a nice 86100C/D mainframe I could switch and keep at least the plug ins. Apart of the trigger, there is not much interesting in the mainframe. It's just a windows computer. The 54750 has somehow the smell of UCSD-Pascal. More a belly feeling, I cannot prove that.
wrt eye diagrams: they can tell you a lot of what is wrong with your laser driver. Excessive bias for example leads to an overshoot at the beginning of the bit cell. The Bias pumps a lot of electrons in the laser diode to an elevated state, and when lasing really starts, they all go off immediately. After that event, the optical level collapses to that what is re-pumped on the fly. You can never see that in simulation unless you include the laser diode physics. (Yes, that's simplified. The diode is never completely off.)
\Gerhard
Last edited:
The box was (is?) an Anritsu 1763 that may have had a problem. They are also huge. We would take that and the Tek DSA8200 on the road to impress tv sales guys. it was ultimately pretty silly but it did sell lots of cables. And at least it wasn't magic fairy dust. I'll try to find out what happened to it. It was all on the marketing budget. Engineering could barely afford used computers.