Author: Ed

Tektronix AM503: Q230 Dual JFET Replacement

Some suggested 151-1032-00 replacements obviously won’t work, such as Tekwiki’s 2N5397 single JFET. Bonding a pair into a single heatsink might suffice, but two separate cans generally aren’t identical enough for the purpose.

Curiously, Tekwiki also lists the 2N5911 as a 151-1032-00 replacement, which (being an actual dual JFET) looks more promising. This agrees with another cross-reference, although the “Sim[ilar] to” suggests considerable caution.

The 2N5911 pinout, as taken from its datasheet:

2N5911 Dual JFET pinout

The actual Tek 151-1032-00 can in its heatsink, oriented with the tab at the top (just visible to the right of the heatsink fin):

Tek 151-1032-00 – top view

Testing one side (with the tab on the left):

Tek 151-1032-00 test side A

And the other side (tab still on the left):

Tek 151-1032-00 test side B

A picture being worth a kiloword:

Tek 151-1032-00 – measured pinout

The drain and source over on the left side seem to be swapped compared to the 2N5911, although both gates are on the proper pins. This being a JFET, the source and drain may be electrically identical and it’s possible the tester labelled them backwards. The only way to be sure Tek wasn’t tragically clever is to poke around the PCB to figure out which pins connect to which other components.

So take a picture of the component neighborhood around the Q230 sockets:

PXL_20220105_210538214

Overlay it with a similar picture of the solder side, suitably reversed / recolored / transformed to match:

Tek AM503 – 151-1032-00 area – X-ray traces

The copper-side traces aren’t complete, as the red coloring marks only traces under the soldermask and omits bare solder-coated traces. Some traces on the component side run invisibly under parts. If I were doing it for money, not love, I’d pay more attention to the details.

Devote some time to tracing the traces and labeling the parts:

Tek AM503 – 151-1032-00 area – part IDs

Then doodle out the actual connections:

Tek 151-1032-00 – part connections

R246 shows Q230B lives in the left side of the can, because it’s connected between the B gate and B source pins, and confirms the tester swapped the B source and B drain pins. Whew!

R236 connects the B drain and the A source, confirming the pinout matches the 2N5911.

Comfortingly, the A side gate goes to all those other parts as it should.

So a 2N5911 will drop right into the Q230 socket with the proper pins going to the proper places. Whether it’s electrically Close Enough™ to the Tek spec, whatever it might have been, remains to be seen, but a good transistor circuit won’t depend too much on the actual transistor parameters.

2022-01-07
Tektronix AM503: Noisy Q230 Dual JFET

The fact that changing R220 also changed the noise should have pinpointed the noise source, but such things are always more obvious in retrospect than in real time running. This post should help me start the next debugging spree a bit further up the learning curve.

The AM503 signal path includes a pair of … unique … differential amplifier ICs made by Tektronix back in the early days of integrated circuitry:

Tek AM503 – U370 U350 detail

The picture has the signal flowing right-to-left through U350 and U370, starting with the Q310 dual NPN in the metal can and the Q315/325 PNP pair (both over on the right side near the cable).

The schematic goes left-to-right:

AM503 – Current Probe Amplifier Schematic – Output Amp – Diag 3

This AM503 PCB does not have the change that split Q310 into two separate transistors, as shown in the upper left of the schematic.

The single-ended input signal comes from Q230 at the output of the attenuator on the previous schematic page:

AM503 – Current Probe Amplifier Schematic – Attenuator – Diag 2

Here, U350 mashes the signal together with the DC offset control and makes it differential, U370 filters it, then Q390 produces a single-ended signal for the scope output. The architecture makes sense when you realize the AM503 started out as a mainframe oscilloscope plugin and became a TM500 plugin as the result of a mid-flight product pivot; think of it as the hardware version of technical debt.

Anyhow, the differential output of U370 shows the noise across pins 6 and 8 (yellow and magenta):

Tek AM503 – diff output U370.6-U370.8

Again in retrospect, pins 9 and 5 would have been a better choice.

The white line is the difference between the two pins and resembled the scope output in the bottom trace well enough to satisfy me.

The differential input to U350 on pins 16 and 14 also shows a distinct similarity to the output noise:

Tek AM503 – diff input U350.16-U350.14

It’s essentially impossible to snap a scope probe around those IC pins, but merely extraordinarily difficult to securely grab the tails of the pin sockets extending beyond the solder side of the PCB.

Finally, a look across R317, the emitter resistor between the halves of Q310:

Tek AM503 – diff input – R317

That was enough to finally convince me the problem lay upstream of Q310.

Ruling out the DC level pot required balancing another AM503 atop this one to plug its cable into the PCB, which showed same output noise.

Hat tip to Sherlock Holmes:

“When you have eliminated all which is impossible, then whatever remains, however improbable, must be the truth.”
https://www.goodreads.com/quotes/1196-when-you-have-eliminated-all-which-is-impossible-then-whatever

Because Q230 is socketed, I pulled it out and popped it into another AM503, whereupon the noise followed the transistor:

Tek AM503 – three amps – GND

All three AM503 amps are set to GND / DC LEVEL CAL. The cyan trance is the formerly noisy amp (now with a good Q230), the magenta trace is the formerly good amp (now with the bad Q230) , and the green trace is the best of the three AM503 amps (untouched, for well and good reason) in my collection.

So I must replace a four decade old dual JFET sporting Tek part number 151-1032-00. Such things are available, but in rather grisly as-is condition or as New Old Stock collectibles.

Perhaps a middle ground would suffice? A couple of those should arrive in a while, but it’s not clear they’re a drop-in replacement.

2022-01-06
Tektronix AM503: Baseline Offset Digression

One of my Tek AM503 current probe amplifiers (SN B064098) suffered from DC offsets in the AC / GND / DC modes, to the extent that zeroing the GND (more formally known as “CAL DC LEVEL”) offset wouldn’t keep the other two baselines on the scope screen. Kibitizing with another AM503 owner with a different problem clued me to apply a change made in later units: replace the 1 kΩ resistor at R220 with a 470 kΩ resistor to reduce the source impedance changes between the switch positions:

AM503 – R220 change

For the record, R220 sits parallel to the attenuator shield above and to the right of Q230 (in the black clip-on heatsink):

Tek AM503 – R220 detail

The new resistor somewhat reduced the offset problem, but also dramatically increased the noise level I’d been studiously ignoring, to the point where the AM503 output was unusable:

Tek AM503 – three amps – GND

The rule of thumb is that it’s always a connector or, perhaps, a similar metallic contact in the signal path. The AM503 has a breathtakingly aggressive switched attenuator covering the 94 dB range from 1 mA/div to 50 A/div:

AM503 – Current Probe Amplifier Schematic – Attenuator – Diag 2

The switches are cam-driven bifurcated gold-plated spring fingers contacting gold-plated PCB pads under that aluminum shield:

Tek AM503 – Attenuator Contacts – detail

The spring-loaded thing to the right is R206, the first 50 Ω 2× attenuator in the form of thin-film elements fired on a ceramic substrate. The two switches put C218 into the signal path in AC mode.

You (well, I) clean the fingers by very gently pulling a strip of lens cleaner moistened with isopropyl alcohol through the closed contacts:

Tek AM503 – Attenuator Contact Cleaning

The pale blue cylinder is the attenuator cam roller extending across the PCB behind the front-panel knob. The two switches bypass C218 in DC mode and connect R220 to ground in GND mode.

Clean gold-on-gold contacts are about as good as it gets and those things looked absolutely pristine. After wiping the contact connecting R220 to ground had no effect, it finally penetrated my thick skull that the problem wasn’t in the attenuator contacts and had to be downstream in the amplifier and filter chain.

Reseating all the cable connectors and jostling the (socketed!) semiconductors also had no effect.

Could one of the semiconductors have gone flaky after four decades?

More tomorrow. Spoiler: yup.

2022-01-05
XFCE: Remote Desktop via X11vnc Through an SSH Tunnel
For the first time in a loooong time I (had to) set up remote desktop sharing, starting from an existing SSH login through a single-port pinhole in an immutable router firewall.

The remote PC runs Xubuntu 20.4 LTS and I verified it already had x11vnc installed. If that’s not the case, make it so.

In order to share / control the desktop of a different user (hereinafter known as kay), I must SSH into that PC as kay. My SSH session uses public key authentication and kay has no need for outbound SSH, so just use my PC’s public key in kay‘s authorized_keys file. On the remote PC, where I am signed in as me:
```
cd ~
sudo mkdir /home/kay/.ssh        # kay does not have a public key
sudo cp .ssh/authorized_keys /home/kay/.ssh     # so just copy mine
sudo chown -R kay:kay /home/kay/.ssh     # transfer ownership
sudo chmod go-rwx /home/kay/.ssh     # set proper permissions
```
From my local PC, I can now SSH into the remote PC as kay and start x11vnc through the SSH tunnel:
```
ssh -v kay@remote.address -L 5900:localhost:5900 "x11vnc -display :0 -noxdamage -ncache 10 -ncache_cr -nopw"
```
Still on my PC, aim a VNC client at the local end of the tunnel:
```
novnc localhost:5900
```
Using novnc presents the remote desktop as a web page in a browser, although you may prefer something more traditional.

Somewhat to my surprise, It Just Worked™.
2022-01-03

Blog Summary: 2021

The overall page view count may be down, but people have been replacing water heater anode rods at an increasing pace:

Home page / Archives	23775
Water Heater Anode Rod Access Done Right	8432
CNC 3018-Pro: GRBL Configuration	5301
G-Code and M-Code Grand Master List	5049
Why You Need a 6-Point Socket to Remove a Water Heater Anode Rod	4300
American Standard Elite Kitchen Faucet Disassembly	2621
Toyota Sienna: ABS Trouble Codes	2531
Raspberry Pi: Forcing VNC Display Resolution	2011
CNC 3018-ProXL: Y-axis Extension	1641
Subaru Forester Fuse Boxes	1276
Broom Handle Screw Thread: Replacement Plug	1239
Removing a Water Heater Anode Rod	1221
Auto-V.I.N Gauge Scam	1029
Low Budget Bench Power Supply	984
CNC 3018-Pro: DRV8825 Hack for 1:8 Microstep Mode	980
Kensington Expert Mouse Trackball: Scroll Ring Troubles	929
Whirlpool Water Heater “Lifetime” Warranty: The Good and the Bad	917
Digital Tattoo Power Supply: Polarity Doesn’t Matter	912
Bed Bugs: Thermal Kill	908
Review Phreesia Authorization	834
Huion H610Pro (V2) Tablet vs. Ubuntu 18.04	821
Kenmore 158.17032 Handwheel Clutch Disassembly	799
Shimano SPD Pedals: Creaking Resolved	788
2000 Toyota Sienna: Replacing the Bank 1 Sensor 2 Oxygen Sensor	753
Schwab / Symantec VIP Access vs. Yubikey	751
Toyota Sienna: Rear ABS / Speed Sensor Failure	729
HP-48GX Calculator Disassembly: Case Rivets	688
Reversible Belt Buckle: Post Restaking	683
Kensington Expert Mouse Scroll Ring Fix	629
Makerbot-style Endstop Power Adapter for Protoneer Arduino CNC Shield	616
Displaying Variables in Gnuplot	614
Kohl’s Guest WiFi Terms & Conditions: The Short Version	613
Replacing Phil Wood Hub Bearings	595
Magnesium Water Heater Anode Rod: Seven Years Later	576
Adding a Device to LTSpiceIV	564
Philips Sonicare Essence 5000: Battery Replacement	550
Browning Hi-Power Magazine Dimensions	545
MPCNC: Emergency Stop / Feed Hold / Resume Pendant	544
Raspberry Pi Interrupts vs. Rotary Encoder	538
CNC 3018-Pro: Home Switches	534
Resistance Soldering: Transformer	510
Adafruit Touch-screen TFT LCD Rotation	489
Quick-and-easy IR-passing / Visible-blocking Optical Filter	476
Dis-arming a Steelcase Leap Chair	471
Icecast and Ezstream Configuration	470
Why You Shouldn’t Use Heat Pumps in the Northeast US	465
Baofeng UV-5R Squelch Settings	460
Mini-Lathe Tailstock: Alignment	449
Homebrew Magnetizer-Demagnetizer	446
Arduino Serial Optical Isolator	444
Mysterious Noise in Toyota Sienna Minivan: Fixed!	444
Baofeng UV-5: Squelch Pop Suppression	434
bCNC Probe Camera Calibration	432
Demolition Card GTA 5-10-9	431
Multimeter Range Switch Contacts: Whoops!	425
Realigning Tweezer Tips	421
Schwalbe Marathon Plus and Michelin Protek vs. Glass Chip	418
Kenmore Model 158 Speed Control: Carbon Disk Replacement	417
Kenmore Electric Dryer: Power Resistor Replacement	416
Old Kenmore Sewing Machine Foot Control Repair	414
Closing the Dmesg Audit Firehose	400

Blog Page Views

That adds up to 200 k page views from 122 k visitors, for an average of 1.6 pages / visitor, down slightly from last year. For a variety of reasons, I wrote only 242 posts over the course of the year, so more folks read only the single post matching their search terms.

To give you an idea of how awful online advertising has become, WordPress shoveled 817 k ads at those readers, slightly more than four ads per view. Given the toxicity of online advertising, I just started paying $50/year for a “personal” plan to get a few more gigabytes of media storage, which also let me turn off the ads. Most of you won’t notice, as you already run ad blockers, but it will calm the results for everybody else.

Fortunately, losing the $250 / year income from those ads won’t significantly affect my standard of living.

2022-01-01

Auvon TENS/EMS: Modulation Waveforms

These scope screen shots use the same test setup as the pulse measurements:

Auvon AS8016 – test setup

The sweep speeds run much slower to capture the complete envelope, which can be up to a minute long, with enough left over to show the end of the previous sequence and the start of the next. The Moire patterns come from the scope sampling rate, the display resolution, or changes in the pulse repetition frequency. Blame Siglent for not making the scope’s digital data accessible through the network; screen shots are the best I can do.

The descriptive headings for each screen shot come from The Auvon AS8016 Fine Manual, a PDF version of which you can get from Auvon’s support staff by asking nicely. I identify the modes as Mxx, rather than their Pxx, for reasons that made sense at the time.

Patterns 1 through 16 correspond to the TENS (Transcutaneous Electrical Nerve Stimulation) label and are intended for pain relief / suppression; they should not make your muscles twitch.

P1 – Continuous comfortable tingling.

Continuous 200 µs pulse at 87 Hz

Auvon M01 Pulse

P2 – Comfortable tingling and pulsing sensation.

Continuous 100 µs pulse at 48 Hz

Auvon M02 Mod

P3 – Comfortable rhythmic tingling.

Blocks of 300 ms on/200 ms off, with 150 µs pulse at 48 Hz

Auvon M03 Mod

P4 – Continuous comfortable tingling.

Continuous 100 µs pulse at 48 Hz

Auvon M04 Pulse

P5 – Comfortable and slow tingling firstly, then the frequency is significantly increased, and it becomes a noticeable tingling sensation.

Continuous 250 µs pulses, stepping from 10 to 102 Hz and back down

Auvon M05 Pulse

P6 – Low frequency beating with a slight tingling sensation.

Continuous 250 µs pulses at 2 Hz

Auvon M06 Mod

P7 – Low frequency slight beating firstly and then continuous comfortable tingling.

Bursts of 150 µs pulses for 3 s separated by isolated 200 µs pulses

Auvon M07 Mod

P8 – Low frequency slight beating firstly and then comfortable pulsing sensation.

Bursts of 150 µs pulses for 3 s separated by isolated 200 µs pulses. Seems identical to P7, although the bursts may be slightly different.

Auvon M08 Mod

P9 – Comfortable tingling from shallow to deep with 3-4 seconds pause.

Auvon M09 Mod

P10 – Comfortable pulsing sensation from shallow to deep with 3-4 seconds pause.

Auvon M10 Mod

P11 – Variable comfortable tingling, slight beating and scrapeing [sic] sensation.

Auvon M11 Mod

P12 – Comfortable slight tingling from shallow to deep with 3-4 seconds pause.

Auvon M12 Mod

P13 – Comfortable tingling and pulsing sensation from shallow to more deep with 3-4 seconds pause.

Auvon M13 Mod

P14 – Rhythmic continuous beating.

Much higher voltage pulses!

Auvon M14 Mod

P15 – Rhythmic scrapeing [sic] sensation.

Auvon M15 Mod

P16 – Quick slight beating first, then comfortable tingling.

Auvon M16 Mod

Patterns 17 through 24 sport the EMS (Electrical Muscle Stimulation) label and should make your muscles twitch in various ways.

P17 – Low frequency slight beating.

Continuous 250 µs pulse with idle time at 4.8 Hz.

Auvon M17 Pulse

P18 – Low frequency beating.

Continuous 250 µs pulse at 6.8 Hz.

Auvon M18 Pulse

P19 – Beating from low frequency to a little high frequency.

Auvon M19 Mod

P20 – Muscle twitches at a very low frequency. It feels like a tapping massage.

Continuous 250 µs pulse at 13.5 Hz.

Auvon M20 Pulse

P21 – This program activates the muscle in a short tingling cycle. It is smoother than P1/P2.

Auvon M21 Mod

P22 – This program gently warms up the muscles prior to exercise; it feels like a rhythmic massage. Increase intensity until you get a strong but comfortable muscle movement.

Auvon M22 Mod

P23 – This program uses a pulse frequency appropriate to fast twitching muscle fibers. It improves their anaerobic capacity and is used for improving maximum muscle strength.

Auvon M23 Mod

P24 – This program gently warms up the muscles prior to exercise; it feels like a rhythmic beating and comfortable tingling. Increase intensity until you get a strong but comfortable muscle movement.

Auvon M24 Mod

Despite the icons on the unit’s display, the manual suggests you can apply pretty nearly any pattern to any muscle, but now we all know what’s coming out of those jacks …

2021-12-30
Auvon TENS/EMS: Pulse Waveforms

The Auvon AS8016 TENS/EMS unit produces bipolar pulses with no net DC offset, so the UI controls the negative and positive amplitudes equally. The range has 20 steps, with the screen shots here set to 10 units. The actual output voltage depends on the mode, with some modes producing a peak voltage well above the others at the same UI setting.

It’s worth noting the effect comes from current passed through skin and muscle, rather than voltage applied to it. The test setup uses a 500 Ω resistance to make the current vary linearly with the voltage (which is definitely not the case with human bodies): a 20 V pulse passes 40 mA through the resistor:

Auvon AS8016 – test setup

The simplest bipolar pulses always start with the negative phase. The shortest pulse width is 100 µs:

Auvon M02 Pulse

And 150 µs:

Auvon M03 Pulse

And 200 µs:

Auvon M01 Pulse

Up to 250 µs:

Auvon M06 Pulse

Some modes have a short zero-voltage pause between the negative and positive phases:

Auvon M17 Pulse

The pause can be the same duration as the negative and positive phases:

Auvon M14 Pulse

Some modes have pulses starting with the positive phase, others switch the leading phase during the course of the output modulation.

My casual survey of the consumer-grade field suggests the pulse waveform has less to do with well-tested effects and more to do with marketing or straight-up woo, but I admit to being a cynic.

2021-12-29