Test Equipment Anonymous (TEA) group therapy thread

[mansaxel]

Now whether you like the "Direct/Reflecting" design concept or not... that's a whole 'nuther argument. 

But that is the same argument. Several time/space disjunct sources for the same sound is a time smearing hell. The idea is fundamentally broken.

[Peter_O]

OK. Back from audio land to real TEA stuff again:

Some months ago I rescued this litte guy.
Now I want to bring it up to speed.



The friction wheel is missing at the frequency dial.
I'm searching for clear pictures. Best I could find so far is the video of Curious Marc:
[url=https://youtu.be/4TFSBH4AEmg?t=797]]https://youtu.be/4TFSBH4AEmg?t=797]https://youtu.be/4TFSBH4AEmg?t=797

From That it seems to be just a disk with some kind of rubber at the front that presses against the scale wheel from behind.
I plan to print a wheel, cut a rubber disk from 'something' and glue it.

Any better ideas or better pictures?

Looks similar to the setup on my 334A Distortion Analyzer. I'll take some pictures later.

They are moulded onto the shaft, with a spring pushing it against the main dial, this is from my 333A distortion anaylzer, the one with the smashed meter.


Could also have a look at the 3310A here, that one is probably slightly smaller than the wheel on the 333A.

David

Any better ideas or better pictures?

Those are useful. I used mine in VCO mode to simulate part of a 4-20mA device.

Try plumbing washers or o-rings.

Note that the rubber washer is behind the wheel. Make sure both knobs spin freely, or the washer won't have enough friction to move the main control.

(see same quote above)

That is basically correct.

It is a hard plastic, as shown below and in factory's picture. The "ledge" looks as if it aligns with the edge of the dial, but I'm not convinced of that, neither am I convinced it is important.

Internally a bronze leaf spring gently pushes the knob's stem outwards to provide gentle pressure against the dial. The light pressure ensures it slips easily when the main dial is rotated. That might not work as well if it was "sticky" rubber.

Looking at your picture, it might be worth considering seeing if a thick rubber tap washer would fit into the knurled recess visible in your stem, e.g. https://www.amazon.co.uk/Assortment-Rubber-Washers-Basin-Shower/dp/B07Q3WZQHK

Any better ideas or better pictures?

Two pics of my 3310A are attached. The plastic is the same as the plastic of the knob itself. No rubber. The shape is slightly conical and the complete axis is spring-loaded to push the disk of the fine tuning knob against the dial.

Some pictures of the fine tuning mech from the hp 3310A function generator, also used on 3310B, the shaft has the plastic disc moulded on to it, the same as the control knobs and was probably made the same way.



1 The top surface is angled, the one spec pictured has worn down quite a bit to have a flat on the edge.
2 For comparison, the shaft with moulded disc from 3310A on the left and the larger one from the 200CD tube oscillator on the right.
3 A failure mode of these (200CD), it has cracked, will probably fall apart same as the controls do.
4 And finally, the complete fine tuning spring mech from the 207A sweep oscillator.
(Peter_O: Pictures rearranged to save space)

David

A big thank you to all of you! (Hope I haven't overlooked anybody, thank's to you too!) and for taking the effort!
Highly appreciated!

For other readers with the same problem: Do follow back the links to the original posts of the guys above, because some of them contain more pictures attached!

The shaft of my 3310A is running smoothly and the spring pressure seems to be fine too.
With all your high quality input I'll start my experiments and will come back here to report.

Thx a lot again!

That worked out fine.

I printed a small disk with a 10deg cone on top, 1.2 mm layer height, 100% infill, PLA for now, and drilled the hole to precisely 5.8mm.
Drilled a 5.8mm hole into a scap piece of plywood and put some alu tape on top.
Stick the shaft into the plywood.



Put some CA gel onto the shaft and press the disk into position.
Pull shaft and disk to some distance to the plywood/alu and let dry.

Some smoothing on the lathe.



And there it is with some grease already.



Works perfectly and allows for quite a decent frequency adjustment, 5Mhz here.



OK, the dial disk is somewhat off, but adjusting and calibrating will be the next step.

[med6753]

OMG, this thread is degrading into an audiophool mud slinging contest.  I will speak no more audio. 

Analog rules, digital drools. 

[TERRA Operative]

Post details of the mods you make to the TDS210 comms module to make it do the maths function here, I'm interested in that as well, might well do the same to mine then as well, please.

I've done it past fall, based on this post/tread:
https://www.eevblog.com/forum/testgear/need-tektronix-tds220-firmware-file/msg1383618/#msg1383618
As the IC's:
RAM: AS6C1008-55SIN
EPROM: AT27C040-90JU
Mouser has them in stock.

Can you do me a favour? I assume you used the file 'TDS2MM_163-0986-00.zip' in that thread to flash the image?
Can you boot your scope and tell me what version that image is? It should say near the top of the boot screen.

I have 163-0986-02 which is v1.04 in my other TDS2MM module, which I'll rip and upload to Tekwiki when my TL866II arrives, but I'm interested in what version 163-0986-00 is.

Also, what version firmware is your scope? I don't suppose you'd be up for ripping the firmware if it is stored on an EPROM instead of the later flash memory too?

[Saskia]

found a Picoscope 2205 at my fav distributor ...  pristine condition (you could say: new).
Could have gotten some Uni-T stuff at significant discount but decided against it. Still too expensive to flip ...

[Zoli]

Post details of the mods you make to the TDS210 comms module to make it do the maths function here, I'm interested in that as well, might well do the same to mine then as well, please.

I've done it past fall, based on this post/tread:
https://www.eevblog.com/forum/testgear/need-tektronix-tds220-firmware-file/msg1383618/#msg1383618
As the IC's:
RAM: AS6C1008-55SIN
EPROM: AT27C040-90JU
Mouser has them in stock.

Can you do me a favour? I assume you used the file 'TDS2MM_163-0986-00.zip' in that thread to flash the image?
Can you boot your scope and tell me what version that image is? It should say near the top of the boot screen.

I have 163-0986-02 which is v1.04 in my other TDS2MM module, which I'll rip and upload to Tekwiki when my TL866II arrives, but I'm interested in what version 163-0986-00 is.

Also, what version firmware is your scope? I don't suppose you'd be up for ripping the firmware if it is stored on an EPROM instead of the later flash memory too?

Scope updated to V1.19, since I've bought 2 EPROM's, just in case; TDS2MM_163-0986-00 is v1.00. To get the screen: Utility->System Status->Misc.

[tggzzz]

I hate to break it to you, but the first thing your ears do is a Fourier analysis in the cochlear, then your nerves translate that into a set of digital PCM signals and finally your brain tries to reconstruct the signal using a neural network to comprehend it. So any arguments that the whole natural signal chain is analogue fall apart.

Quite.

Add in that the whole ear-brain system is horribly non-linear in every dimension you can think of. If it wasn't, audio compression wouldn't work!

Perhaps so. But has anyone determined our hearing's sample rate?

You'll have to define what you mean by that.

For example, if you consider the long (by electronic standards) response time of the nervous system, you could conclude that (say) 100 samples per second would be sufficient to cover it.

Bet it's infinitely better than any digital format. After all, billions of years of evolution.

In the same way that the human eye is better than a camera? NOT!

[mnementh]

Now whether you like the "Direct/Reflecting" design concept or not... that's a whole 'nuther argument. 

But that is the same argument. Several time/space disjunct sources for the same sound is a time smearing hell. The idea is fundamentally broken.

Awesome. Then snip a wire to the extra tweets on your system. Who cares?      Or, as you say, refuse to touch the brand.

I still like the way my 301s sound as rear surround speakers. My space, my ears, my sound preference.

I also like the EQ tweaked to boost highs and lows with a flat middle. As I get older, that just gets worse; I don't hear those frequencies as well, and I've always liked to feel the bass.

None of that has anything to do with lumping everything they've ever made in together with the 901 and derivative designs and stating flatly that they were all the same "big pile of 4in speakers" design. Come on.



Today I saw a pair of AVID 101s at the Thrift; sadly, severe water/mildew damage to the bottom areas of the enclosure. 

That kid who usedta build speakers for fun is still in here somewhere; he is screaming at me to go back and get them just to save the tweets and woofers from certain doom... 

mnem

[mnementh]

I hate to break it to you, but the first thing your ears do is a Fourier analysis in the cochlear, then your nerves translate that into a set of digital PCM signals and finally your brain tries to reconstruct the signal using a neural network to comprehend it. So any arguments that the whole natural signal chain is analogue fall apart.

Quite.

Add in that the whole ear-brain system is horribly non-linear in every dimension you can think of. If it wasn't, audio compression wouldn't work!

Perhaps so. But has anyone determined our hearing's sample rate?

You'll have to define what you mean by that.

For example, if you consider the long (by electronic standards) response time of the nervous system, you could conclude that (say) 100 samples per second would be sufficient to cover it.

Mmmmhmmm... but we're talking about ~15,000 cilia in the human ear, all processed in parallel. x2. And I for one have no idea what the bit depth per single cili (?) would be; certainly much more than binary.

Bet it's infinitely better than any digital format. After all, billions of years of evolution.

In the same way that the human eye is better than a camera? NOT!

It is obvious that the human eye is superior to a camera at what it is intended to detect; otherwise we wouldn't have all these publicized problems with Apple's child-safe bunk identifying a digitized pic of a dog as probable child-porn. It's both the sensor and the processing that makes a working eye vs a working camera.

The difference with a camera is how much more range it can detect and of course the fact of once again, parallel bus for the eye vs serial for a camera.

For a real appraisal, you cannot separate the human eye from the visual cortex... it is a system all directly tied into the brain itself. The problems we have quantifying how it all works are similar to that of trying to gauge sound with a tape measure; the tools we have at our disposal are... ahem... less than ideal.

mnem
 

[Saskia]

spent the evening fighting a bloody Fritz 3000 repeater and found out that a slave fritz box could have done the WLAN repeater job via powerlan bridge a hell of a lot better (i.e. works) than the dedicated repeater which dedicatedly does not work.
Should have bought ubiquity. Awaiting payment for the appartment. As soon as that's here:
make a 10k down payment on my mortgage
order fibre
oder photovoltaics (I want LG / Enphase ...) with battery backup
look into independent water supply
look into additional power sources ...
look into alternative heating solutions (such as feeding surplus solar stuff into crypto mining rig and reuse the heat to heat the house, sumthang like that ...

[bd139]

Brand name of the day 

[Robert763]

found a Picoscope 2205 at my fav distributor ...  pristine condition (you could say: new).
Could have gotten some Uni-T stuff at significant discount but decided against it. Still too expensive to flip ...

I'm a Picoscope fanboy. Don't forget the serial data decoding. Pico Technology include as standard what Tek and Keysight chare hundreds for.

[tautech]

Brand name of the day 

Right  and strongly believe you will buy what you had before. 

[factory]

On turntables, I can suggest bolting the record to an angle grinder and using a stanley knife as the needle to play the god awful circular disks of crap.  At no point do I want to listen to another bloody LP even if it's being played on a machine which floats on Jesus' farts.

Photon did half of that, some years ago. 


David

[Cerebus]

I hate to break it to you, but the first thing your ears do is a Fourier analysis in the cochlear, then your nerves translate that into a set of digital PCM signals and finally your brain tries to reconstruct the signal using a neural network to comprehend it. So any arguments that the whole natural signal chain is analogue fall apart.

Quite.

Add in that the whole ear-brain system is horribly non-linear in every dimension you can think of. If it wasn't, audio compression wouldn't work!

Perhaps so. But has anyone determined our hearing's sample rate? Bet it's infinitely better than any digital format. After all, billions of years of evolution.

The simplistic answer for sound duration would be around 170 Hz - it takes around 6ms for a nerve to form an action potential and recover enough to be able to be triggered again. Fortunately a lot of parallelism is involved. Hearing samples in the frequency domain and the sampling bandwidth is both non-linear and overlapping. I doubt that it's an accident that we perceive two successive frequency doublings as perceptually linear (i.e. the same musical pitch one and two octaves apart). The average human can just detect a pitch difference of 0.0035 (e.g. at 1kHz a 3.5Hz pitch difference, at 5kHz a 17.5 Hz difference and so on).

[Saskia]

that fritz repeater is a piece of shit not even worth the 50 grams. I'll try to return it.
Using the second fritz box as a repeater works like a charm.

[factory]

Oops... I got carried away.... no I am not going to move on to sorting the ICs. Just remembered I had a couple bins of more trannystors to sort first ! 
TO3 and TO220 packages. some salvaged and usable (long enough legs), but mostly intact legs so must be old stock.
At least these packages are easier to grab and to read the part number from, than all these pesky TO92 I just sorted.

OK so I am done with the TO3 packages now. Started with the TO3 as much less of them than the TO220 packages... low hanging fruits first ! 

23 of them in total. 3 that I could not identify, if people can help, you are most welcome. If not... straight to the junk bin, since an unidentified anything is as useful as a rock. Maybe I could saw them open out of curiosity...

Anyway as you can see they are :

- S-1828 made by " SI " in 1974. Or so I understand.

Then a couple made by RCA :

- CVT.......... 418
- CXK ....... 3M

So I am left with 20 that I could identify. Not always easily. Some don't have datasheets that Google can find me, so I have to rely on databases that give basic specs for old stuff. Problem is these sites don't always agree... and by a long shot. One gave a collector current of 15A max and the other pretended it was 45A.... 

I have a BD142-7 but the datasheet doesn't say anything about any suffixes, so don't know if the -7 suffix has any importance... I fear it has but will never know what it means 

...snip...

Well the S-1828 gives nothing, but 1110715 cross references to NFR8117 diode on the parttarget NSN stock list website and searching for NFR8117 finds a Digital parts list here http://vtda.org/docs/computing/DEC/Handbooks/EB-21850-75_SparesKitHandbook_1982.pdf
The DEC book lists it as PIV 75, 20A in T03 package, under part# 11-10715-00 (also cross-references back to the NSN list). Does it test as a diode?


The RCA parts are 64595, those five digit codes are usually custom part codes, finding data for equivalents seems to be a problem, except for some 40xxx parts, which are listed in the 1967 RCA databook I have here.

David

[factory]

I am just done sorting all these transistors in these 2 plastic boxes I showed. Old man's stock from ex-work.

A chore, looking at each and every of these tiny TO92  package one by one to see what they were.

Then pulled all the datasheets.

Created a folder named "Stock", created a spreadsheet in it, below, to list all the components, write down type, basic specs, comments, quantity.... and created subfolders to hold all these datasheet for future reference. I am getting organized....

Somehow I find it satisfying to look at this spreadsheet. I now know at a glance everything that was in these boxes, specs and quantities. Information is power.... 

I didn't find any BU508A nor a BD139, but I do have a BD136 so if you flip the last digit upside down, that gives me BD139 

....snip....

Then the big surprise was this AD142 and AD149 (one of each) TO3 PNP !!
They are...  GERMANIUM !!!   Daddy retired 20 years ago, not 60 years ago !! I have no idea why the hell he had these two antiques in his drawers at work !
But well, maybe I can make use of them in my Vintage TE restorations, one day.. though I guess if I opened the cans I would be greeted with lots of tin whiskers 

It's usually the AF11x series (AF117 for example) from Mullard that often suffer from tin whiskers. Some EU made ones supposedly have epoxy type fill, instead of silicone jelly and the whiskers don't grow through epoxy. Of course Ge trannystors can fail in other ways.

Recommended reading if you haven't seen it, NASA report on tin whiskers in AF11x/OC17x type Ge transistors.
https://nepp.nasa.gov/whisker/anecdote/af114-transistor/2005-Brusse-tin-whiskers-AF114-transistors.pdf

I've got some BD139 but no BU508A, as shown in an old post below, where I was having trouble trying to sub a hp custom divider IC in a 5248M counter, which turned out not to be the problem anyway, ref oven required a tweak. 

How to sub transistors like a pro:

2N1132   -> fuck it BD139 BD140 will do

2N1711  -> fuck it BD139 will do.

 

Fixed that for you. 

Bugger it. Thought I had put bd140 down for first one. I looked at it and though 2n2905A first but then decided bd140 and wrote bd139. Apparently, I suck! 

Apparently I made a mistake too when I said I hadn't got any 2N1711, just went to look at the NOS spares for another TR and noticed I have some 2N1711 in one of the drawers, but failed to spot them in the spreadsheet.

Anyway still having more problems with the 5248M, this time the A26 5MHz to 10MHz multiplier & 1MHz divider board, not getting the 1MHz out. Think these may sort the problem out with the 1820-0082 IC, but not sure if that's enough to sub for the IC. 

David

[bd139]

Right  and strongly believe you will buy what you had before. 

When I have moved, maybe something a little more upmarket from the same vendor of course 

[bd139]

On turntables, I can suggest bolting the record to an angle grinder and using a stanley knife as the needle to play the god awful circular disks of crap.  At no point do I want to listen to another bloody LP even if it's being played on a machine which floats on Jesus' farts.

Photon did half of that, some years ago. 
David

Oh that's brilliant that. Hadn't seen that one 

[BU508A]

that fritz repeater is a piece of shit not even worth the 50 grams. I'll try to return it.
Using the second fritz box as a repeater works like a charm.

I've used in several installations AVM repeaters, including the 3000.
No problems at all, they just simply work. Installation can be a little tricky
but once they are appearing in the mesh, everything is fine.

[mnementh]

Brand name of the day     

So... when does it arrive...? 

mnem
 

[Saskia]

that fritz repeater is a piece of shit not even worth the 50 grams. I'll try to return it.
Using the second fritz box as a repeater works like a charm.

I've used in several installations AVM repeaters, including the 3000.
No problems at all, they just simply work. Installation can be a little tricky
but once they are appearing in the mesh, everything is fine.

it trashed my config a couple of times, grabbed mesh master role and completely fucked up my previously working mesh config plus Internet connectivity forcing me into crawlspaces to physically rewire and reset stuff.
I fucking hate it. my life is too short to fart around with this  Hope Reichelt takes it back.

[tggzzz]

I hate to break it to you, but the first thing your ears do is a Fourier analysis in the cochlear, then your nerves translate that into a set of digital PCM signals and finally your brain tries to reconstruct the signal using a neural network to comprehend it. So any arguments that the whole natural signal chain is analogue fall apart.

Quite.

Add in that the whole ear-brain system is horribly non-linear in every dimension you can think of. If it wasn't, audio compression wouldn't work!

Perhaps so. But has anyone determined our hearing's sample rate?

You'll have to define what you mean by that.

For example, if you consider the long (by electronic standards) response time of the nervous system, you could conclude that (say) 100 samples per second would be sufficient to cover it.

Mmmmhmmm... but we're talking about ~15,000 cilia in the human ear, all processed in parallel. x2. And I for one have no idea what the bit depth per single cili (?) would be; certainly much more than binary.

You appear to be doing the equivalent of confusing baud rate with bit rate, or clock rate with instructions per second.

Bet it's infinitely better than any digital format. After all, billions of years of evolution.

In the same way that the human eye is better than a camera? NOT!

It is obvious that the human eye is superior to a camera at what it is intended to detect;

The human eye/brain is crap, both absolutely and relative to other independently-evolved eye variants.  Start by considering why our eyes have a blind spot. Other eyes have avoided that grossly bad engineering. Then go on to consider the remarkably small field of decent vision, i.e. the fovea centralis.

We can't perceive just how crap, since we are limited to what we can percieve. Well, that's not quite true, as shown by optical illusions. Have a look at the 148(!) examples at  https://michaelbach.de/ot/index.html Many are familiar, but some are astounding; this one is gobsmacking https://michaelbach.de/ot/mot-mib/index.html

otherwise we wouldn't have all these publicized problems with Apple's child-safe bunk identifying a digitized pic of a dog as probable child-porn. It's both the sensor and the processing that makes a working eye vs a working camera.

That, and Teslas, poor performance don't make the human eye/brain good.

The difference with a camera is how much more range it can detect and of course the fact of once again, parallel bus for the eye vs serial for a camera.

Too simplistic, but at least you are acknowledging that cameras are better in many ways.

For a real appraisal, you cannot separate the human eye from the visual cortex... it is a system all directly tied into the brain itself. The problems we have quantifying how it all works are similar to that of trying to gauge sound with a tape measure; the tools we have at our disposal are... ahem... less than ideal.

Indeed you can't separate them; the brain does a surprisingly good job of guessing/inventing what the eye omits.

Analogies are dangerous, and often confuse the issues.

[mnementh]

On the (Woodworking) Bench Tonight:



So a few days ago, I bought these two belt sanders (see here in the What'd Ya Buy? thread), figuring I'd do a Amazon vs Horror Fraught cheapies shootout; see which was the best. 

Turns out... they're both the same exact fucking sander under the plastic shell. Aside from the part that's red on the Bauer, everything... motor, electronics, belt drive to gearbox, even the drive & shoe casting, etc... all exactly the effing same Chinesium bits.

The only real difference then is the features, aside from the fact the dust collector on the Bauer is 50% blocked due to a design flaw, and doesn't work worth a fuck. The Galax Pro has adjustable front handle, a dust collector that works, and that little flip-up cast cover so you can actually do some inside curves with the thing. 

So... I just took the red one back to Horror Fraught and used the refund to get a jitterbug and sanding discs. 

I've made a lot more progress than you can see here... I've finished the stripping and now working on smoothing it out again. I'm going out to work on it some more, now that the kiddles are fed.

mnem
*toddles off to get his grit on*