I have a bad probe interface card from a 4 channel unit. Since the label at the front is nearly impossible the transfer, and the screws was put on before the label covers it, plus the FFC connector is impossibly tight even for hot air to get there without much damage, it’s near impossible to save it with a 2 channel card.
Out of curiosity, I removed the label sheet to see what’s inside it:
The PCB is the same for 2 channels or 4 channels. The 2 channel version simply have channel 3 and the aux trigger hole covered (channel 4 is the external trigger port in a 2 channel model). So technically, you can the the excess label and cover up the “Ext Trig” text, but it won’t look professional. If it’s your personal unit, then feel free to go with the hack.
54810 series (first generation) Infiniium uses the same PCB for 2 channels and 4 channel models. They slap on a different rubber keypad sheet and the button labels depending on whether it’s a 2 channel or 4 channel model.
I received a 4-channel front panel keypad module that was ruined by ripped pads around the relay while trying to replace it. Instead of trying to fix it, I transferred the rotary encoders to a 2 channel PCB which are in abundance, and I noticed this:
It seems like HP/Agilent at some point tried to save a few pennies by skipping the SMD grains (resistors, capacitors, inductors, transistors, diodes) surrounding Channel 3 and 4 for the newer board on the right.
So if you are looking to repair a 4 channel front panel keypad with 2 channel PCB, you should preferentially select ones from the older lot which has all the parts for 4 channels populated except the rotary encoders. If not, time to get a pair of SMD hot tweezers and transfer the grains one by one.
It’s basic (signal processing) mathematics that square waves (or any waveform with sharp edges) carry harmonics that doesn’t die fast enough, therefore not band-limited. The more sudden the transition is, the more harmonics you need to preserve to faithfully represent a signal in practice.
Unless the signal itself is known to be band-limited (like sine waves, classical modulation schemes, SRRC), it puts the burden on the test instruments involved to provide large bandwidths and the matching high sampling rate (needs to do better than Nyquist).
The technology today provides generous bandwidths and sampling rates for oscilloscopes at reasonable prices, but synthesized function/waveform generators with the same bandwidth/sampling rate can easily cost 10 times if not even more. For the money to buy a used not-too-old 500Mhz pulse generator, you can buy a 4GHz 20GS/s oscilloscope!
For oscilloscopes, users are often aware that the rounded square waves/transitions they see on the screen is due to bandwidth limitations, and will account for the reality distortion in their mind. If your square wave clocks are pushing the sampling rate of the scope and the combined bandwidth between the scope and the probes, you should very well expect that you cannot catch much glitches and pretty much use the oscilloscope as a frequency counter or check for modulations.
For synthesized signal generators, bandwidth and samplers are way more precious. Oscilloscopes generally has 8-bit vertical resolution (256 steps), but synthesizers typically has 12-bit vertical resolution (4096 steps) or better. There are imperfect techniques trading vertical resolution and sampling rate, but there is no free lunch.
I have on my bench these 12-bit function generators:
an old Analogic 2045B (roughly US$1300, 400Mhz, 800MS/s, 2Mpts) and
a much more modern Agilent/HP 33120A (roughly US$600, 15Mhz, 40MS/s, 16Kpts)
that I’d like to illustrate the value of getting a higher bandwidth (and therefore higher sampling rate) unit. The 33120A has a much finer control over frequency/amplitude/offset steps (2045B only allows fixed point increments) and might have better noise characteristics and much smaller form factor considering Analogic 2045B is made in the 1970s and HP 33120A is made at least 20 years after that. I would have liked to keep only my 33120A or 33255B on my bench to save space, but once you’ve seen this screenshot, you’ll know why I’m still willing to cough up the space for 2045B:
The upper waveform (Channel 1, yellow) is a from Analogic 2045B while the lower waveform (Channel 2, green) is from HP 33120A. They are both set at around 15Mhz and you can see that when approaching the limit of the synthesizer, 33120A rounds off the square wave to almost a sinusoid. Square waves gets ugly quickly above 10Mhz and this is as far as 33120A is capable of.
On the other hand, the square waves are bang-on for 2045B, and is still decent at around 50Mhz (my BNC cable starts to come in question). That’s why it’s still worth getting a high bandwidth synthesized function generator even if your budget only allows for clumsy old models if you use your function generator for something more than sinusoids.
Note that function generators are typically designed to pump out to 50Ω loads and the amplitude displayed in the function generator () assumed so. That’s why beginners gets confused why they read when the function generators says : the function generator sends out nearly double the voltage so the potential divider formed between the 50Ω output impedance of the function generator and the 50Ω load impedance will split the voltage into half. If you set your oscilloscope to 1MΩ load, you are getting , which is nearly .
More importantly, if the load is not matched, the small capacitances in the chain will distort the waveform received by your oscilloscope severely at higher frequencies, so you can barely get a square wave at fundamental frequency above 2Mhz undistorted if you feed it into an oscilloscope with 1MΩ input impedance, while in reality the signal generator, cables, and oscilloscope can do much better.
Most cheap low bandwidth oscilloscopes (like 100Mhz) do not have 50Ω option. Nonetheless the impedance needs to be matched if you work with square waves at 2Mhz or above. Just buy a 50Ω feed-through BNC ‘termination’ adapter and plug it right at the 1MΩ input port. In reality, it’s a divider between 50Ω and (1MΩ//50Ω), with the scope seeing of . For all practical purposes the oscilloscope sees or .
With 50Ω feed-through BNC ‘termination’ adapters, make sure you work out the impedance matching if you split the signals if it’s not the simple nearly 1:1 potential divider halving the voltage. At low frequencies, the amplitudes will be off, but when you start going into Mhz range and above, your signal will be distorted badly as well.
When old equipment’s fail, they do fail in waves, depending on the failure modes induced by the original design. Last week when I turned on a TDS 784A in my inventory check, something smelled bad and the display was garbled (it has displays, but straight lines turned into wiggles).
I already replaced the caps for the processor board, keyboard and RS-232/Parallel Port module preventatively and the unit used to work fine. So it boils down to either the power module or the CRT driver.
Despite it’s unlikely to be the power module (didn’t feel any fan speed changes, display brightness changes, or hiccups in power), I used my nose to make sure there’s no burnt electrolyte smell from the power module. Indeed there wasn’t.
Sniffing can be a very valuable tool to repairs. The smell came from only one narrow area of the board so I limited it to 3 capacitors next to each other:
I took them out and cleaned the PCB and noticed that the wipes has a bit of green and black stuff on it. That’s how I can tell a capacitor just peed all over itself. The culprit is C321 and C323.
Note that the component layout for this color CRT driver, 678-1402-07 (the board has silkscreen saying 671-2373-389-1344-01) does not match the component locator I have with my TDS 544A schematics. Nonetheless, it’s nearby if you look around.
Just to confirm the capacitors I took out are the culprit, I used an LCZ meter as an overkill ESR tester to test them:
ESR for these two caps should be at the order or milli-Ohms if they were any good. I took the one next to the two offending capacitors out to test it, and the ESR looked OK so I put it back. The true reason is that I don’t have that capacitor value on hand at the time of writing, but that also helps to narrow down the true cause.
I replaced these two capacitor and the display worked correctly (not garbled). The brightness is a little bit high which can be adjusted down.
The next problem is that the shutter color changes out of sync back and forth a slow then fast rate till it gets stable after warming up for a while. I did a lot of troubleshooting, changed a bunch of capacitors and transistors and shutter board, but no avail. In the process, I smelled electrolyte evaporating with the flux and I decided to give the board a full wash with dishwasher detergent and waterpik (then dry the big part with a leaf blower, spray with 99% rubbing alcohol to the water out and left it dry). Bingo!
Lessons learned: do not leave the electrolyte leaks on the board even if it’s an old fashioned single-sided through-hole with relatively simple thick traces. I thought it’s not going to matter until I see visible corrosion, but I was wrong. Could it be the electrolytes left on the board forming weakly conductive paths that disappears when the unit warms up (the electrolyte dries up)?
In the process of replacing all the electrolytic capacitors on the board, I smelled fumes mixed with electrolytes in some areas (other than the two above). However, I didn’t record it because I measured the ESR for each capacitor that I pulled and compared to the ones I’m about to put in.
In addition to the two capacitors mentioned early in this post, here are few capacitors that the ESR of a new part is significantly lower, which might be first places to consider replacing before recapping the entire board. They are all measured at 1V, 1kHz:
Nonetheless, the only useful technique that contributed to this board being repaired is finding out where the smell comes from. The rest (reading at schematics, measuring voltages, checking waveforms on an oscilloscope, swapping out parts) are all red herrings.
In the process of troubleshooting with schematics, I also noticed that the schematic for the old TDS 544A color CRT driver is actually pretty much the same (including component numbering) as this newer board 678-1402-07 while I was troubleshooting with it.
Looks like the component layout was slightly shifted to make room a different batch of flyback transformers (there’s a riser board for the flyback transformer in 678-1402-07 used in 754A/784A). Although the component locator sheet isn’t exact, the components are within 1 inch of what’s found in TDS 544A CRT driver’s component locator. No biggie if you don’t have the schematic for the newer color CRT drivers. Just look around and pay attention to the silkscreen. Common sense will lead you to the right part.
I was repairing a HP 3560A that does not start up at all. The system is very modular that there are only 3 main units: LCD/keyboard/DSP board (A1), Main Processor board (A2) which manages power sources and contains the backlight inverter as well, and the analog section (A3).
Obviously the first thing to look for is where the power is managed, which is the main processor board (A2). I bought another unit (with a different defect, i.e. it boots) hoping to use it as a reference but to my surprise, the A2 board is slightly different! In fact, the LT1120 voltage regulator that I’m seeing unusual pulses in V_out pin (should be flat) is not even there!
I’ll save the repair story for another post, but for information preservation purposes, I took pictures of the two different revisions of the A2 assembly from 3 units, shown below:
New: 03569-66502 Rev B
Old: 03560-66502 Rev B, Rev C
The top part of the board is essentially the same. The new board uses Toshiba TC551001CP-85L while the old board uses Sony CXK581001P-70L for Static RAM. They are likely pin compatible and it’s just availability differences. (Ignore the randomly placed caps at the bottom of the new A2 board. I desoldered them to test if they are good).
If you pay close attention, at the top of the new revision board, it has an unpopulated DIP-8 socket and an extra 74HC174N chip, and an extra digital I/O port at the mid-bottom left edge below the SRAM. They are reserved for 3569A (a better version with a noise tracking generator) as it uses the same board. The old revision A2 board works for 3560A only. It’s a topic for another post.
The bottom part is quite different. The classical 4-diode full-bridge rectifier on the new board is not shown at the main side of the circuit board for the old A2 assembly. The new board looked much denser.
Most importantly, the old board uses MAX666 as the 5V voltage regulator while the new board uses LT1120. The pinouts are different and chip features varies a little bit, so the power management section is not topologically identical.
There are no components on the back side of the new revision A2 board (they are both supposed to be single sided PCBs), but two diodes are squeezed in at the back of the old revision A2 board, and there’s an extra resistor flyover:
My suspicion is that the diodes are intentional as there are specific through-holes for them (most like they other half of the bridge rectifier), but the resistor is an after-spin rework.
Finally, for information preservation, I also took pictures of the old A2 board (03560-66502 Rev B) from another unit I have:
If you consider the relative ease of use for less computer savvy people, 3560A/3569A is versatile yet designed specifically for the most commonly used measurements in acoustic and mechanical vibrations. It’s still excellent value for what it offers despite it’s a made decades ago since it doesn’t overwhelm you with convoluted choices so you can gets the job done once you’ve setup your recipe.
HP has designed the unit very well, with the exception of the LCD screen which cannot be found on earth, all through-hole components on single layer board and well organized structure and silkscreen makes servicing a pleasure.
I can repair and rebuild 3560A / 3569A with new battery pack and clock battery. The hardest problem to track down is no-boot, and the hardest surgery to make is the backlight. I also know how to talk to the unit from modern computers as well, so data capture is not a problem. Call me at 949-682-8145 for consultation.
If you have an original HP 3560A sitting for years and haven’t changed the battery pack yet, it’s guaranted to be dead. Here’s the shell of the battery pack 1420-0584:
Just to give you no hope attempting to reused the battery pack, I dissected one of them and show you how much of a disaster inside it:
This pack has been sitting for so long that the cathode (+) wire is severely corroded (not all of them are like this). Even the connector turned green. I cut open the wires and gave it a tap, and a bunch of copper oxide bits falls off.
Now that I acquired the tools, parts and practice to make one from scratch (I used to need the old battery pack). It’s not cheap (given the tools, research effort, and most importantly, small quantity), but I can custom build them for you if you don’t want to deal with the hassle and steep learning curve.
The first pack is the most expensive, additional ones are much cheaper because of the reduced overhead. Lead time is around 1~2 weeks unless you pay extra for me to express order the ingredients.
You can call me at 949-682-8145 or just go to humgar.com.
One of the most memorable things about xkcd’s (in)famous circuit diagram is the 666 timer (at the top right corner of the comic), a parody of the famous 555 timer:
Today I was analyzing the logic board for HP 3560A and noticed some batches uses a MAX666 chip and it immediately reminded me of this xkcd joke. Turns out Maxim Integrated makes a voltage regulator with a cool part number! I bet the engineers back in the days must have started with 666, giggling uncontrollably, and added 665 and 664 to see if it can get past the regulators or the censors (puns intended for both).
I received an Agilent N9340B 3Ghz Handheld Spectrum Analyzer with a note that it passes all self-tests but does not respond to input signals. I took the gamble that it’s the RF input connector got disconnected somehow.
I opened up the case and noticed that the 40Mhz cable was unplugged, so I was half-correct. I connected it and got a signal at the precise frequency, but the amplitude doesn’t look quite right. It’s around -20dB off. When I scanned it across the full 3GHz band, I noticed the amplitude roll-off when I scan below 800Mhz, and I got very little signal left when I get to somewhere near 10Mhz.
I tried running a user calibration with a 50Mhz CW source but it failed amplitude calibration. Apparently the unit is not fully working. No self-test errors though.
So I opened up the unit and the RF section. The front side of the board doesn’t have any visible signs or unusual smells, so I suspected the improper gains is caused by the input attenuator HMC307:
I was about to order the chip, but because of the lead time, I decided to just take a picture of everything and analyze it off-line:
After removing the screws holding the N-type terminal so I can get to the back side of the board for taking pictures, I noticed the RF out connector just fell off the board with the pad:
That means the RF out is not touching the board. I never would have suspected that it’s the problem, since this unit does not have the tracking generator option enabled and hence the RF out port is pointless. But for the sake of completeness, I resoldered the connection after I put the board and the connectors back to the RF module slab. It’s ready to be stowed away for another day when the attenuator chip arrive.
Guess what? Once I put unit back together, I turned it on again and everything works perfectly! The power level is flat and within 1dB of what my 8648C pumps out. I did the user amplitude calibration again, it passed, and everything was spot on!
My suspicion is that the path gap created a capacitor between the board and the type-N terminal, which messes up the termination and created reflections. The bottom section of this picture is where RF out meets RF in:
I didn’t study RF as my EE speciality, so if you are a RF design expert, please let me know what you think the reason might be in the comments section.
For TDS 500~800 series, a batch of CRT driver boards, color and mono, regardless of how heavily they are used, have bad flyback transformers. After turning the unit on continuously for half a day, the screen might stretch and disappear.
If you have a matching CRT driver board with a CRT tube, I recommend instead of swapping the CRT driver (seemed more straightforward), extract the flyback transformer from the donor board instead. The reason is that the adjustments needed from replacing the flyback transformer is far less than re-tuning a different CRT driver board to match the tube.
It’s impossible to tune the CRT driver board while it is in the case, since the processor board covers it during operation (unless you have special cables for the Acq/Proc interface to replace the interconnect PCB card), it’s done ex-vivo like this:
I bought a ribbon cable extender and built a 2-pin jumper extender by salvaging them from CRT driver boards with toasted flyback transformers:
The first thing to check for is the +21V which is used to generate many voltages across the board (pun intended here): it affects brightness, scale, offset and linearity everywhere. If there’s any adjustments to be made, this need to be done first.
This voltage can be tapped by hooking the positive (red) lead to the center (output) pin of LM317 (3-pin linear regulator) at U90. If you have an alligator clip instead of a grabber, you can also hook it up to ‘pin 4’, which is the body of the regulator.
You can pick many spots for the ground pin. Since I’m using a grabber, I’d pick another big 3-pin IC sitting on a heatsink for the ground lead. In this case, it’s Q10, the transistor that drives the flyback transformer. It’s the pin nearest to the short edge of the board (behind the red lead, sorry):
Here’s a picture of blank board showing how many trimpots are there:
Only the brightness and contrast dials are documented in the service manual. The rest, I had to locate them in the schematic one by one. Before that, I kind of figured out most of them by trial-and-error but had a few of them wrong, especially the voltages (there are three: +21V, screen and HV adj.): they all have the same effect. There are also some more obscure trims like center focus and horizontal focus (variable inductor). Now I know exactly what each dial does.
It’s hell of a lot of work to figure this out. I have some new old stock CRT straight from Tektronix at Beaverton, and it’s the reserve to support customers who bought color TDS 500~800 units from me. Almost all used units out there have problems (or going to have problems soon), and so far I’m the only one selling units with 1 year warranty (extendable to 3 years for extra).
If your unit is not under warranty included when you bought from me, and want one of these new color CRT tubes with shutter, I’ll almost require you to send your unit to me for installation unless you can guarantee that you can figure it out without my help. It’s $500 full-service with the tube included. Call me at 949-682-8145.
One day I was working on an old first generation Agilent Infiniium oscilloscope (namely 54810A) while I was too tired and rushed, I accidentally short-circuited the main acquisition board because I forgot to reconnect the probe-compensation port pin back to the acquisition board after taking out the front-panel and putting it back: the jumper with exposed metal (the heat-shrink over it was a little short the way Agilent manufactured them) swiped over something and I heard a loud bang, the scope shuts off, and I smelled the magic smoke.
My heart sank. Just saving a few extra minutes being careful checking everything (despite I opened and closed those front panel ten dozen times) I thought I lost an expensive PCB that’s almost the cost of the whole oscilloscope, and even if I can fix it, it might drain me at least a week.
Given that I know a short fried something (there’s a bad smell). I didn’t even bother to turn the unit on again until I’ve located what fried. Turning something that you know it’s fried on again just risks further damage as it can load other parts of circuits.
Following the smell, I found a burnt IC on the other side of the board (needs to be painfully disassembled as the front panel /w BNC nuts needs to be taken out all over again), and I looked up the part number: ST L6201. It’s sitting on Channel 1’s front end section between the attenuator relay block and the ADC hybrid and there are 2 of them per channel.
Given the location of the component, it’s clearly the L6201 populated at Ch3’s slot is not used since it’s a 2 channel oscilloscope. So I transferred the chips to replace the broken ones at Ch1:
What is an H-bridge driver (L6201), that’s supposed to control motors, doing in an oscilloscope, especially the front-end section? I googled “H-bridge driver in oscilloscope” and nothing relevant turned up. Then I went back and read a little more on how an H-bridge driver is really used. L6201 is a DMOS Full Bridge Driver with four power MOSFETs (switches) that basically sends current through an inductive load (typically motor) that might have stored energy (momentum) that might need to be drained (brake) to stop faster.
Turns out it’s a slick way to drive mechanical relays in the input attenuator given the amount of due care needed to accurately manage the current demand and switch transitions. There are 3 relays in the attenuator module. I suspect up to two relays can be switched with each L6201 by placing a diode in serial with with a relay coil, and repeat it (in parallel) with the diode reversed in the other branch. Is it an overkill? Let me know in the comments section.
I also noticed, while dealing with the main acquisition board for 54615B/54616B, there is a L293D chip under the attenuator block shield for each channel. It’s a 4 channel driver with half-bridge, also intended to drive inductive loads. This one is more explicit about being used to drive relay solenoids as well as motors. So this is nothing new; It’s just not too many people talked about using it on oscilloscope architectures.