If the error code is E5622, the most common one, the fix is to replace the sram things in the bottom. That's easy.
formatting link
I have maybe 10 of the mainframes and a couple dozen heads, worth a fortune at the original price. Beautiful instruments. My personal scope is an old b+w 11802, which takes nicer pics than the later color versions. It will internal trigger, too, with the delay lines.
Anybody interested in picosecond stuff should get one of these. They do weigh 50 lbs or so.
--
John Larkin Highland Technology, Inc
picosecond timing precision measurement
jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com
Change the circuitry to have balanced Q/_Q drive to two otherwise identical in layout parts and look at the outputs. Betcha you see decidedly different timings. Then swap the layouts and note that difference essentially does not change...
This thing was a 32-channel "discriminator" for nuclear signals. It had two comparators and a scheme to compare a delayed and attenuated signal with the original, to achieve a measure of rise-time compensation. Called, in that field, a "constant fraction discriminator". So, each of the 32 sections had a power island with caps and a series resistor for decoupling. I just measured the dip in voltage when it switched, and was astounded to see a huge dip in voltage. Yes, there are all SORTS of parasitic inductances in the decoupling caps that makes real quantitative numbers pretty suspect.
Yes, quite likely the voltage inside the chips was massively affected, if I saw that much dip at the outside.
Oh, I assumed that any attempt to measure it had an error bar of about +/-
50% or worse. But, it told me I had FOUND the problem, and I just had to find chips that didn't have such horrible shoot through.
Hmm, you're right, the waveform on page 7 is labeled CP Input (
formatting link
).
Definitely nothing in that data sheet to suggest transition times in the hundreds of ps. It would be funny if it turned out that cheap carbon film resistors act like shock lines.
I'm using a nice LVDS->CMOS receiver chip that has typical rise and fall times in that range, according to a datasheet plot of t_R vs C_load. (I'm not at my computer or I'd post a datasheet link.)
Are LVDS outputs balanced? AIUI "differential" means positive and negative voltage, but "balanced" means equal and opposite current. So if the swing is 0 to 5 on each line then it might be balanced if the source and sink currents are equal, but differential means the swing is, say -5 to 5. If those definitions are right, then what's the purpose of differential signals that aren't balanced?
The output structure is usually drawn as a current steering pair, but that's a lie: there's a fairly low common mode impedance in there. A necessity, in fact!
Engineers so often follow, blindly, into the belief that differential = magic solution. But no, you can't fool the signal. Exceed the input CM range and your signal goes bye-bye!
Having a high-ish CM output impedance, does two things:
It passively DC biases the output, so Vcm = Vdd/2.
It provides some damping for the transmission line: as opposed to a perfect CCS, which will permit high-Q resonances on the line (making the system very sensitive to AC upset).
The CM impedance being higher than most CM transmission line impedances, isn't ideal for the most common use case (LVDS pairs on PCB have a pretty low Zcm, maybe 50 ohms), but does help with worst-case "accidents" (you should never run a non-isolated pair through space alone, but if you do -- for example by using a ribbon twist cable -- the high Zcm of that condition will be terminated even better).
Note also that, since the source is a high impedance, and the load(s) are high impedance (there's no CM termination in the standard), an LVDS bus looks like a 1/2 wave resonator, so at least the first resonant mode falls at a higher frequency than if it were driven with a very low impedance logic driver (which would make a 1/4 wave resonator).
And yes, the output pin pair is carefully synchronized, so that little common mode AC voltage is transmitted -- the currents match.
LVDS (and compatible standards) I think is the only case where Vcm is allowed to float, at a fairly high CM compliance.
Other differential standards nail down Vcm firmly: RS-422/485 transmitters are simply pairs of logic buffers, one complemented (422's outputs are always on, 485 is tristateable; 485 also has diodes in series with its output transistors, so it won't clamp line voltage to the rails, and can be powered down without loading the line). The Z_CM of a 422/485 transmitter is quite low indeed (~20 ohms?), which can give rise to standing waves during transmission.
Or Ethernet PHYs, which are just open drain pins, actually; they require source termination resistors, and the isolation transformer is obligatory for success (and EMI reduction!) over any distance.
Or CAN PHYs, which drive the terminated line through diodes (like an RS-485 transmitter, but instead of a full H-bridge, it's one diagonal pair, so the "negative" line state is 0V instead of negative), so Vcm can be wide if undriven (wide noise margin), but is firmly nailed to Vdd/2 during transmit (which can cause standing waves).
It's a FIN1002. Fairchild are being poopyheads and encrypting their datasheets for some reason, but you can get it from AllDatasheet: . Check out Figure 19.
Cheers
Phil Hobbs
--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC
Optics, Electro-optics, Photonics, Analog Electronics
160 North State Road #203
Briarcliff Manor NY 10510
hobbs at electrooptical dot net
http://electrooptical.net
The 11802/SD24 has 20 GHz bendwidth, and it looks perfect using the calibrator pulse and its own TDR step. I see about the same waveform using the resistor pickoff vs the 6 GHz HP probe on the other sampling channel.
That 150ish ps rise looks real. Amazing for a 16 cent CMOS part.
--
John Larkin Highland Technology, Inc
lunatic fringe electronics
--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC
Optics, Electro-optics, Photonics, Analog Electronics
160 North State Road #203
Briarcliff Manor NY 10510
hobbs at electrooptical dot net
http://electrooptical.net
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.