Got a 1725a here that needs the vertical IC. It's designation is A5U2 and I THINK the part number is 5061-0024. The manual is a shitty scan, but any o ther numbers I can get out of it are too off the wall, at least that number googles back to SOMETHING.
The Thing about this is that it is missing. Is it a high failure rate part ? Another question I have is if it fits any better scopes. I assume of cour se they would be HP, but why would someone take it out unless either it was bad, or they wanted it to use in another unit. If they used it in another unit, what would that be that would warrant effectively scrapping this one ? A higher end scope using the same part ?
OK, I could see in a business setting where it owuld not be feasible to rep lace the thing even with the same model because of anal regulations, but is that a reasonable assumption ? Also, if they fail that much, maybe I shoul d be leary of a used one. I assume there are no new ones.
According to the print, this thing only has four transistors and a half doz en resitors in it. Think there is a possibility of some sort of retrofit ? Like building the equaivalent out of discrete parts ? Might not be absolute ly perfect but the thing is useless now.
Either that or part it out. Are the other parts worth much ?
The Thing about this is that it is missing. Is it a high failure rate part ? Another question I have is if it fits any better scopes. I assume of course they would be HP, but why would someone take it out unless either it was bad, or they wanted it to use in another unit. If they used it in another unit, what would that be that would warrant effectively scrapping this one ? A higher end scope using the same part ?
OK, I could see in a business setting where it owuld not be feasible to replace the thing even with the same model because of anal regulations, but is that a reasonable assumption ? Also, if they fail that much, maybe I should be leary of a used one. I assume there are no new ones.
According to the print, this thing only has four transistors and a half dozen resitors in it. Think there is a possibility of some sort of retrofit ? Like building the equaivalent out of discrete parts ? Might not be absolutely perfect but the thing is useless now.
Either that or part it out. Are the other parts worth much ?
Thanks. Having the right part number helps no matter what. I notice both of those numbers are lower than 1725, should I assume then that the 1725 is t he best of the "family" ? If that's the case, the mystery remains of why th is one got the axe. Of course the thing might have just went bad, it does h appen.
WOW ! Now that might just save the day. Thanks. I notice a dot RO in the pi cture there, does that mean Romania ? I've been told by some from across th e pond that alot of parts are very hard to get, it's not like the US. They have had to become more resourceful as a result.
So it pretty much boils down to do this unless I can scare up a jubnker wit h the IC. This looks highly doable. the only thing now is the selection of the MOSFETs. Voltage, current and frequency response are obvious, the only question is gain.
It's obvious that the 1K resistors set the gain to some extent, but the hig hest gain possible is not always best in tis type of application. Or is it ? a thousand ohms isn't much, maybe the highest gain within reason would be best. If you want to make a suggestion I am highly suggestable. The main t hing I know is to get devices that will work in the linear region. That wil l probably mean looking at some datasheets because most linear applications I've seen are either audio or video amps and neither of them are going to have the frequaency response for this. At least not necessarily. Also, sure ly the devices exist, it's just that I haven't seen them in the stuff I've worked on.
I actually thought of a retrofit myself, but I was going to stick more to t he original, using bipolar. Of course that was not thiknking outside the bo x in a way - we only need this to do a certain thing, if FETs work, use the m and make it easier. I would guess devices like that were not available wh en they built that unit, but they are now !
Have you ask for a quote from the 100+ companies that won't give you a price online. I found a couple of others looking to replace the same ic in the same scope. You might be better off building it yourself, for reliability. Mikek
You mean the ones you send your email address to and if they ever come up w ith one you have to buy 100,000 of them ? No, I didn't bother doing that.
scope. You might be better off building it yourself, for reliability. "
Not a bad idea really. Any one I get now would be used. The problem now is finding the right MOSFETs. I bet recalibrating it will be loads of fun. Act ually it might not be all that bad. The HF response will probably be better if anything. I might have to remove some of the peaking caps or something if they won't adjust down enough. Then if the gain is really up there of c ourse then I might need to change some resistors. They got the internal 50 ohm changeed to a 47, but that won't mean shit to differential mode. I hope the static plate voltages come close enough ! Then comes getting the thing straight.
OK, at this point I am wondering just what kind of FETs to use for this thi ng. I'm sure y'all are alot more familiar with what kind of MOSFETS are out there than I am. Some of youse iz real injunears, I am just a half decent hack. As such I am not all that familiar with what is available.
Max limits are pretty obvious and not taxing. Just with some quick mental m ath I see Pd as being easy, maybe 3 W if that. Max Id is not even an ampere and Vds about 100 V. The problem is getting the frequency response. Even t he gain should not be all that critical because there is a 1 K resistor fro m gate to drain. The problem is that offhand I just don't know of an otherw ise suitable MOSFET for use up to 275 Mhz. Even a video output in a high en d CRT monitor would not handle it I think.
Because of the massive level of feedback, maybe I could get away with a 100 or 150 Mhz device no ? I doubt it would be a good idea to go below 150 Mhz here, in fact I would like to have 300 Mhz.
So what will amplify that high linearly ? I think a TO 220 case would be ni ce, in fact a plastic TO 220 would be alot better as I won't have to insula te the tabs and there would be less radiation. Altogether much more desirab le I would say.
I haven't had much luck at Digikey. I select my parameters and mostly the t hing comes back with no matches. I really want TO 220 plastic but when I dr opped that from the parameters I got some in a funky RF type cass, round bu t with a football kinda shaped clamp like portion. I would really rather no t use that case style, especially if the drain is internally tied to the ca se. It just looks like a PITA and/or it might make a bit too good of an ant enna. Probably intended for grounded drain operation or perhaps to run in a metal box. I dunno, I just prefer a plastic TO 220 type, unless someone ar ound here has a better idea.
Another thing I haven't done in a very long time, and don't even know if yo u can anymore, is to call a manufacturer and talk to one of their reps abou t it and maybe even get them to send me a couple of free samples. Do they even do that anymore ?
Of course I have no qualms about paying for them, but then again I don't ne ed 5 KW transmitter finals either here. It's a nice scope but not that nice .
Any suggestions ? I am in a quite suggestable mood here.
Well there is enough info there to get the scope working, without looking at Yahoo. The suggested D16NE FET is obsolete. Mouser suggests a
512-HUF75307D3ST, which they then say it is obsolete. These are high current devices, Are deflection plates low impedance? I thought they would be high impedance high voltage devices. If you can't get or don't like these two devices, try looking at other scopes and see if you can find a drive the uses discretes and try their FET/transistor. What is the bandwidth of the scope? Mikek
512-HUF75307D3ST, which they then say it is obsolet"
Thanks, at least I have somewhere to start. I can just look for the spec sh eets on those. I finally did start getting some results on Digikey's select or guide. I guess on soem things it's better to just aplly one filter at a time.
low impedance? I thought they would be high impedance high voltage devices. "
As far as I know deflection plates are very high impedance like a control g rid, though they migh pick up a chargee due to their placement in the envel ope. They do have capacitance though, so at 275 Mhz (the rated -3Db point o f this scope) they could concievably impose some load. As far as DC is conc erned, the load resistors are IIRC like 160 ohms coing out of 53 volts whic h is only about 34 mA. I believe the engineers probably chose tthe 160 ohm value becaause it would make the interelectrode capacitance negligible at t he highest operating frequency. This makes the compensation less of a PITA. there is no other reason to use such a low resistance.
I also notice that on this unit, the current from the load rresistors is pa ssed through two terminals for the deflection plates. In fact it shows it a ctually going through the plates, lke it is not just two pins, it is the tw o ends of each plate. Not sure exactly what they're getting out of doing th at but it doesn't bother me. Of course it must make some difference at 275 Mhz.
This is probably one of the faster non storage CROs out there. In digital t hey can just digitize the waveform and display it at a rate a cheaper defle ction system and amps can handle. Not in this baby, what you see is really what you got. Really, once you get through all the hurdles to get flat band width up to that point, I think some RAM and a processor would be cheaper, but of course not when that thing was built.
Thanks. Later today I am off to look at a few datasheets.
This all a little confusing to me. The suggested mod uses high current FETs (12 or 15 amp). Just doesn't seem correct. Looking at the schematic it shows 165 ohms in series with each plate, with a B+ of 53V, the current would be limited to 320ma.
Here's a schematic I found that makes more sense regarding high voltage low current..
FETs (12 or 15 amp). Just doesn't seem correct. Looking at the schematic it shows 165 ohms in series with each plate, with a B+ of 53V, the current would be limited to 320ma. "
Correct. Actually there is nothing really "in series" with the plates. They operate electrostatically, the only load they impose is capacitive. the wa y they have it congigured is apparently they have electrodes at each end of the plates, for whatever reason. I do not see much advantage in it. the pl ates are not all that large, and 270 Mhz's wavelength is about 1 meter.
The schematic to which you linked is not what this thing uses. It already h as pushpull drive coming in. The transistors that drive it have their bases directly connected and the emitters are fed. Maybe, MAYBE the chip only pu ts out a single ended signal, but I really doubt it. Even looking back to t he delay line, which is literally a roll of coax, it isa actually to rolls of coax. That would seem to indicate the signal is already fully diffeentia l at that point.
However people there and at places like tektronix are anal like that. Of co urse they have their reasons, some of them have to do with rejecting interf erence, and some of them have to do with avoiding the generation of interfe rence.
I just thought 2100 pF was a bit much, but then across a 1K resistor and fe eding a 160 or so ohm resistor, it is not all that bad.
Incidentally, that schematic in your link makes a mistake that is one of my pet peeves. the fact is that it is incorrect to call a phase inersion 180 degrees out of phase. A phase shift is when you delay something. If you del
gh an inverting amplifier it is inverted, not SHIFTED in phase. Delayed by
180 degrees and inverted are only the same on waveforms that are bidirectio nally symmetrical, such as a sinbe wave, or a 50/50 duty cycly square wave or ene a pefect triangle wave. Only things like that which only contain odd order harmonics, it does not apply to any waveforms that are non-symmetric al BOTH WAYS.
But I didn't come to rant about that. The point is that in most scopes, the phase splitter doe not work like that, which is like ones in old tube audi o amps, but not even all of them. Some of them actually used a tube that wa s high gain and configured about like an inverting OPAMP. Two equal resisto rs, the junction of which is at the grid of the invertor stage resulting in unity gain. Most scope amps don't even do that, they just make either the emitters or bases of a pair of transitors common and feed one of the eleman t while bypassing the other. this causes it to invert, and after a coule of stages of the the CMR is high enough. (common mode reduction, not rejectio n ratio)
If you have good enough transistors, one stage does it well enough, but in these instruments, after the first phase splitter, ALL stages are different ial and they ALL have some CMRR. It all adds up, and you get a perfectly ba lanced signal.
Interestingly, they think this has been reinvented with the way the signal gets to the display on some TVs now, called LVDS. The same scheme is used i n the USB, HDMI, firewire and other signal transmission methods.
It is about the same as the old twinlead anrenna wire back in the old days, just refined. It doesn't radiate alot and it rejects noise. by keeping the signal perfectly balanced and at a low lever, capacitive effects of the tr ansmission line are also minimized allowing for greater bandwidth.
But it is still a rehash of the shit that came out befor eI was born, and t hat was 1960. Circuit boards and IC were still forthcoming back then.
So they haven't really invented much, they have just refined it some. boils down to, for one, you are right, the current needed is not that high, but then if it happens to have the good rating then it does. I guess it will no t blow as easily. I found that in Digikey's selector guide. They are talkin g 12 and 16 amps, I don't need anything like that even if there is a friggi n capacitor across those plates. thing is I was looking at TO220 type cases because I believe they will make it alot easier to build the thing. there is only so much room for mounting and I don't want to leave them hanging by their leads which might only be .040" in diameter, and then maybe use sili cone to keep things from shorting out.
Anyway, back to the drawing board on finding a MOSFET. One thing I do have to do is ggo to Electronic Shithouse (ESI in Mentor). REALLY, I mean no dis respect calling them that, my Uncle came up with the name. He used to go th ere every Saturday for quite some time. That is a long story. His kids (my cousins) built their own stereos.
Anyway, I tried to go to their site and it was down at the time. I will try again and see. Since the scope is old enough for a military pension, maybe the parts shouls be older as well. However the MOSFETS will not be as opld beecause they did not exist then.
You know how I know that ? Because HP would have used them ! lol. Those eng ineers have no clue of costing, they are trained that way. If the thing cos ts ten grand, so what ?
OK, enough chit chat for now. I am offf to do something. No idear what but. ......
1) The original circuit used bipolar NPN transistors in the hybrid. You might consider using some high voltage video driver transistors for your design.
2) The circuit is not driving the deflection plates directly. They are driving a delay line with taps that connect to the deflection plates. This is so the deflection signal follows, in time, the electron beam as it moves through the deflection plate assembly.
3) HP really spent a lot of time in getting the whole system to work correctly. Making changes that try to second guess what was done may lead you down a rabbit hole.
Good luck with getting it to work. They really are a nice scope when right.
I have HP1725A schematics from two different sites and both have a 165 ohm resistors in series with each plate. Then I find another 50 ohms in series with the transistor emitters (the ones in your defective IC)
The way they have it configured is apparently they have electrodes at each end of the plates, for whatever reason.
One electrode goes to the transistor that pulls it low, the other electrode goes to the 165ohm resistor that is tied to +53V. Yes it is push/pull.
Post a schematic or something to convince me the schematic I have is not correct.
btw, the 165 ohm resistors seem to reside on their own board (A5A1) all by them selves.
Mikek
PS. I have seen some owners raving about their HP1725A scopes.
No, the delay line for the deflection plates is inside the tube, parallel to the segmented plates. Its purpose is to align the instantaneous deflection voltage with the moving electron beam. Just for the learning experience why don't you calculate the speed of the beam as it passes from the beginning to the end of the vertical deflection plates. remember it is in a vacuum i.e. ideal.
The delay line you speak of is usually a coil of coax and does exactly what you said. It lets you see the beginning of an event after the trigger.
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.