Test Equipment Anonymous (TEA) group therapy thread

[bd139]

Very standard design akin to Thurlby / TTi and it is of the same heritage.

Feels a lot like diving in the better models of 80s -90s audio gear from companies like C-Audio, BSS or Klark-Teknik; they usually had a bit more bracing to be tour-proof, but the general look is a lot like that. Usually nothing fancy; if the 'phools only knew how many TL072 stages their "Reference" recordings have passed through they'd self-ignite.  Soundcraft usually had a discrete pair of BC550 / 560 level transistors at the input, but after that it was all opamps.

Nowt wrong with a TL072. Yeah I'm a fairly big fan of the TTi/Thurlby/Black Star etc etc stuff. It's not fancy, it's not beautiful but what it is roughly translates to very sound engineering. It's easy to maintain, well documented, cost effective and fit for purpose. Which is what you mostly want out of something.

[bd139]

Not 100% certain, but I think that they were the same company, just a different brand name, they certainly shared the same address so that's a pretty good clue, as did Thandar and then I think they decided it was better to run with a single brand and called themselves TTi Ltd.   

I might drop their support line an email and ask them. The history of AIM-TTi is a bit of a weird one so would be good to iron it out for the sake of actually being able to quote something authoritative on the matter. Will add it to my list of things to do and see if I can find out.

[bd139]

WIFI with async web services NICE TO HAVE, but I do not want to chose a wifi board which compromise the stability.

Learning MQTT right now, very interesting stuff, but I have no experience with it. Pondering to have a stupid (BUT REALIABLE) wifi board (looks like not ESP32) and move the horse power into the MQTT brocker/server.

Slightly off topic question, why it is so hard to find a RELIABLE SBC platform?

The clue to doing reliable, scalable, maintainable software systems is to have a good separation between functional parts. A minimal sensor implementation that throws MQTT messages, to a hub pair, off of which consumers, several, can siphon the data is a clever thing. One of the consumers should be persistent storage, others can be things like near-real-time displays et c.

I've done a lot of data collection in the times before MQTT et al, where I got the values in, made SQL statements of them, threw them on the end of a defined file name, and then put that file as standard-in to a mysql client, with && rm file on the end. If the SQL operation runs well, the file is cleaned, if not, it is a cache of statements to be inserted later. This very well survives shaky Internet connections.

Wretching at the thought of putting data inside MySQL but anyway... I have a similar solution deployed that uses SQLite on the client as an inbound and outbound transactional queue. That synchronises over HTTP to the server end by sending message batched in JSON where it is stuffed on an ESB. Works completely offline or online and is reliable. The client end can actually be preloaded with data via an initial pull which will populate the local SQLite.

Only about 100k lines of C# that. But if you change something on a WPF data model form it will automatically queue the updates with no code needing to be written.

Code: [Select]

[
 { "type": "client.update", "serial": 8199, "payload": { "id": 232, "name": "Bob" }},
 { "type": "client.update", "serial": 8200, "payload": { "id": 232, "name": "Bobette" }}
]

And then Apple invented exactly the same fucking thing with Core Data 

(perhaps I should be paid more)

[nfmax]

Borrow the one from your meter or microwave oven or meat thermometer.

mnem
 

What "meter" ? No not going to dismantle my poor u-wave oven !! 
No meat thermometer here either.

I want to design this thing, the time pressure / emergency makes it even more fun...

So... you don't have a proper bench thermometer, or even a "K" thermocouple as used on almost every DMM since the dawn of time...?

What. The. Actual. Fuck. 

Seriously, this is an egregious shortcoming we shall have to pick on you for... at least the next 5 minutes. 

If you go dismantle that cheap digital thermometer you have on the wall right now we'll forgive you... for now. 

mnem
pssst... on most microwaves that have a thermometer, it plugs in. Easy-peasy to borrow.

A whole new world of Kelvin-nuttery awaits! There is nothing to beat the satisfaction of making up an ice point in a 1 litre vacuum flask (using shredded ice from de-ionized water, naturally, for best accuracy), immersing the lab reference PRT probe, and having it read a stable 0.00˚C for hours. My second-best PRT probe read -0.02˚C. A random K-type thermocouple probe read +1.2˚C, which is about average for a K-type. T-types are rather better, usually about ±0.5˚C, but not guaranteed.

Thermistors are generally quite stable, but the response curves vary quite widely and are often mismatched with the curve assumed by the meter. With individual calibration they can give excellent results, within ±0.1˚C.

[bd139]

There is FreeBSD for Raspberry Pi.

mnem
"It's just like tryin' to fuck a alligator: A whole lot of thrashing about,
and even if you're successful, no good can possibly come of it." ~me

That is a perfect representation of the Raspberry Poo. It's not a computer; it's a bad idea which is really badly implemented wrapped in a layer of marketing under tight control by a rabid fanbase and moderation on their official forums. Eben and Liz can literally fuck right off. Putting FreeBSD on it is a folly at best when there are real, proper, non shite computers out there you could throw your efforts at. Even that cranky ass 12 year old spooge coated shitbox you got from a house clearance for nothing because the guy who was clearing it out didn't want to touch it incase he got Herpes is a better computer. It's also 100x less likely to fuck up on you even if it's limping along like a rabid squirrel and you have to scrape the jelly like layer of nicotine off it.

[Vince]

Yeah !!!

555 wired up, oscillates !

Put a couple 100K resistor and 1uF cap as an easy starting point with round numbers. Ran the math, gave a theoritical frequency of 5Hz, just the ball park I need so impeccable.

Works perfectly. ON time is spot on the OFF time as it ought to be, and frequency spot on 5Hz.  I am in heaven, it's a good start ! 

I calculated that a Frequency of about 20Hz (18.2Hz), would be good. Divided by 2^16 (x2 8 bit counter TTL chips cascaded), it would give me a one hour period.

So let's now fine tune this to get about 20Hz and add a pot to vary the duty cycle 

[mnementh]

So you mean they aren't bad ones, I can just buy the first random generic nu branded super cheap one on Ebay or XYZ, and it will be just fine ?
3 bucks at that ?!

Well OK will buy one real soon then ! 

The MX54C can use either a PT100 or PT1000 thermocouple, if I believe the documentation I have, below.

PT100 & PT1000 sensors are thermistor-type devices; they are also similarly standardized as a type K thermocouple. Both are available for $2-8 shipped on fleaBay; the PT100s are more sensitive to EMI, so usually have shorter shielded leads and may have the shield pulled out to a separate pigtail. These are most often used in industrial applications, so will usually be either unterminated or terminated in fork-type screw terminals. You'll have to make up some banana ends on them.

I must be losing it... more than usual... I've looked that pdf over 3 times, and I can't see where it says anything about using the PT100/1000 sensors.

mnem
*toddles off to nuke a power controller*

[bd139]

Yeah !!!

55 wired up, oscillates !

Put a couple 100K resistor and 1uF cap as an easy starting point with round numbers. Ran the math, gave a theoritical frequency of 5Hz, just the ball park I need so impeccable.

Works perfectly. ON time is spot on the OFF time as it ought to be, and frequency spot on 5Hz.  I am in heaven, it's a good start ! 

I calculated that a Frequency of about 20Hz (18.2Hz), would be good. Divided by 2^16 (x2 8 bit counter TTL chips cascaded), it would give me a one hour period.

So let's now fine tune this to get about 20Hz and add a pot to vary the duty cycle 

No RPN! But TI-85. Acceptable 

Actually really nice calculators and not the same heritage as the TI-83 as you can tell because the people who designed this didn't make the exponent key shifted. The screen was shit but they fixed that in the TI-86, which I had one of 

[nfmax]

WIFI with async web services NICE TO HAVE, but I do not want to chose a wifi board which compromise the stability.

Learning MQTT right now, very interesting stuff, but I have no experience with it. Pondering to have a stupid (BUT REALIABLE) wifi board (looks like not ESP32) and move the horse power into the MQTT brocker/server.

Slightly off topic question, why it is so hard to find a RELIABLE SBC platform?

The clue to doing reliable, scalable, maintainable software systems is to have a good separation between functional parts. A minimal sensor implementation that throws MQTT messages, to a hub pair, off of which consumers, several, can siphon the data is a clever thing. One of the consumers should be persistent storage, others can be things like near-real-time displays et c.

I've done a lot of data collection in the times before MQTT et al, where I got the values in, made SQL statements of them, threw them on the end of a defined file name, and then put that file as standard-in to a mysql client, with && rm file on the end. If the SQL operation runs well, the file is cleaned, if not, it is a cache of statements to be inserted later. This very well survives shaky Internet connections.

I've played with MQTT a bit myself, running mosquitto on a pi, but never used it in anger. However, the latest firmware upgrade for my FireBrick router/firewall adds a built-in MQTT broker, so I might give it another go. The broker does not have to run on the same system as the database server, and there are good reasons to keep them separate.

[cyclin_al]

Nothing in the modern vehicle is worse than this and if it can't be easily disabled I'm buying something else ! 

I see, so dying in the modern vehicle is less of a nuisance?

You need get out and about more. 
Recently as a passenger in a 4x4 .....something that you city folk may not be aware of is they are frequently used off public roads and at combatively low speed where also as a rural work vehicle in and out of it often opening gates and such a permanently bleeping beeping seatbelt warning all day everyday is nothing but misery !

I'll stick to my guns on this, if it can't be disabled it can go back to its maker and be shoved where the sun don't shine !   
Door not shut buzzers get the same treatment here too......oh inside light is On or the door warning light on the dashboard and I must have an audible warning too ? FFS if I can't see the obvious door open indicators then I shouldn't be driving. 

In the case of a work vehicle used for work, I have to agree.

Try hitching up wagons or trailers.  Such a PITA!!!
Need to reverse 3 inches to line up properly?
Okay, jump in the truck and put it in reverse. 
Not allowed due to seatbelt not on.
Not allowed due to door open (but I need the door open to judge exactly 3 inches distance over the ground).
Do all those things that the software insists on according to the truck design and it turns into a WAG.
Try anyways.
Undo all steps (seatbelt, door, etc.) to then be able to get out of the truck and go look.   Oops, went 1 inch too far.
Repeat all steps again, and this time end up 4 inches away instead of the original 3.
Repeat many times until lucky (or all karma has been paid back). 

As for road vehicles, I somewhat like how the seatbelt helps hold oneself firmly in place for spirited driving. 

[nfmax]

So you mean they aren't bad ones, I can just buy the first random generic nu branded super cheap one on Ebay or XYZ, and it will be just fine ?
3 bucks at that ?!

Well OK will buy one real soon then ! 

The MX54C can use either a PT100 or PT1000 thermocouple, if I believe the documentation I have, below.

PT100 & PT1000 sensors are thermistors; they are also similarly standardized as a type K thermocouple. Both are available for $2-8 shipped on fleaBay; the PT100s are more sensitive to EMI, so usually have shorter shielded leads and may have the shield pulled out to a separate pigtail. These are most often used in industrial applications, so will usually be either unterminated or terminated in fork-type screw terminals. You'll have to make up some banana ends on them.

I must be losing it... more than usual... I've looked that pdf over 3 times, and I can't see where it says anything about using the PT100/1000 sensors.

mnem
*toddles off to nuke a power controller*

PT100 & PT1000 (the number refers to the resistance in ohms of the sensor at 0˚C) are metallic resistance thermometers, not thermistors, which are semiconductors. Being low impedance, they are not particularly susceptible to EMI but you do have to take precautions against the effects of lead resistance. PT100s are normally supplied with 4-wire Kelvin (naturally) connections. These are often shielded, but not always.

The resistance of a platinum thermometer increases about 0.4% for each ˚C, so a high-resolution, 4-wire resistance measurement is all you need. Plus some maths, as the response curve is non-linear. There are two parts to the error: the shape of the response curve, which depends only on the purity & heat treatment of the platinum wire; and the deviation from nominal resistance at 0˚C, which depends on the accuracy & repeatability of manufacture. The specification for these sensors appears in IEC 60751

Generally speaking, better accuracy means higher cost. You can start at a few pounds for a IEC 60751 class C thermometer, going up to thousands for an individually-calibrated probe with an accuracy in the 10's of mK level over a wide temperature range. That's without venturing into the realms of metrology-grade SPRTs...

[tautech]

As for road vehicles, I somewhat like how the seatbelt helps hold oneself firmly in place for spirited driving. 

Not those on the average runabout, for a proper track session you need a 5 point to stay in the same place in the seat to get the best performance from your wheels......much like the lap belt on my tractor when you're pushing things to the limit of what's capable.......one error and you're toast......pushing near the edge is always fun. 

[Neomys Sapiens]

Of course, the newly aquired 'Pocket' multimeters (Pokemults?) had to undergo the DMMcheck test!
Here the results:

Benning MM-T3

VDC    4.99V
100R    100.1Ohm
1kOhm   0.998kOhm
10kOhm   09.99kOhm
100kOhm   100.0kOhm

VAC(100Hz)   05.51V
Frequency   100.0Hz
Duty Cycle   49.9%

VAC (10kHz)   2.533V
Frequency   10.00kHz
Duty Cycle    49.8%

1nF      1.33nF
10nF      10.8nF
100nF      100.2nF
1µF      0.998µF   

Current: no current measurement

Fluke 115

VDC    4.999V
100R    100.1Ohm
1kOhm   1.000kOhm
10kOhm   09.99kOhm
100kOhm   100.0kOhm

VAC(100Hz)   4.987V
Frequency   100.0Hz

VAC (10kHz)   2.73x..2.76xV (a bit unstable)
Frequency   10.00kHz

1nF      1nF
10nF      10nF
100nF   101nF
1µF      1003nF

DC current   0.001A
AC Current   0.000A (all frequencies)

[Vince]

So you mean they aren't bad ones, I can just buy the first random generic nu branded super cheap one on Ebay or XYZ, and it will be just fine ?
3 bucks at that ?!

Well OK will buy one real soon then ! 

The MX54C can use either a PT100 or PT1000 thermocouple, if I believe the documentation I have, below.

I must be losing it... more than usual... I've looked that pdf over 3 times, and I can't see where it says anything about using the PT100/1000 sensors.

mnem
*toddles off to nuke a power controller*

Eh ?! WTF...... Oh, I see !!!  Looks like I attached the wonrg file, I attached " Metrix MX 50 Series.pdf " instead of " Metrix MX 50C Series.pdf " , my bad sorry ! 

It's on the 3rd page, it shows a table listing / comparing the specs of all the models in the range. MX54C is third column. Temperature stuff is near the bottom.

I corrected my post, and also attached the correct PDF here as well so you don't have to scroll back. See how nice I am ? I would do anything to seek Dwagon's forgiveness 

[factory]

Edit to keep on topic since we've advanced a page....



I popped the cover off of the 4800A and took a few more pictures of it.  Normal HP construction of the era - well laid out and easy to access, no surprises.
Top view of open chassis:


The oscillator circuitry resides inside the shielded box in the front right corner:


Other pics at: https://pmanning.smugmug.com/Electronics/HP-4800A-Vector-Impedance-Meter  (Crappy iPad pics, at some point I'll take better ones with the DSLR)

-Pat

Great post. Thoroughly enjoyed that. I actually spent an hour or so reading the HP 4800A service manual and reverse engineering the A1 / A9 utility amplifiers. Very nice little discrete opamp.

Now I want one 

Good luck finding one for a sensible price, tried buying the same one twice now, at a certain auction house that likes to bid on their own items, instead of using a proper reserve. 
Also looks the same size as the 5245 series counter, really you probably don't want one. 

David

[srb1954]

Are those PCBs only tin coated or is just the lighting/camera that make them look that way?

On my 4800A the PCBs are fully gold plated as was normal for HP equipment of that era.

They're tinned - no gold on them.


Mine is a 1968 engineering rev (816 serial number prefix), and based on a quick survey of component date codes appears to be at least early 1969 for its birthdate.  It looks like the model first appeared in the 1965 catalog.  What year or engineering rev is yours?

-Pat

Serial prefix 1246A, which would suggest a 1972 build date.

However, looking at the date codes on the reed relays in the input module indicates that it was built around 1976 so that may have been replaced at some stage. The serial number label on the module is still a 1246A prefix but it looks like it was peeled off and stuck down again.

[Cubdriver]

Edit to keep on topic since we've advanced a page....



I popped the cover off of the 4800A and took a few more pictures of it.  Normal HP construction of the era - well laid out and easy to access, no surprises.
Top view of open chassis:


The oscillator circuitry resides inside the shielded box in the front right corner:


Other pics at: https://pmanning.smugmug.com/Electronics/HP-4800A-Vector-Impedance-Meter  (Crappy iPad pics, at some point I'll take better ones with the DSLR)

-Pat

Great post. Thoroughly enjoyed that. I actually spent an hour or so reading the HP 4800A service manual and reverse engineering the A1 / A9 utility amplifiers. Very nice little discrete opamp.

Now I want one 

Good luck finding one for a sensible price, tried buying the same one twice now, at a certain auction house that likes to bid on their own items, instead of using a proper reserve. 
Also looks the same size as the 5245 series counter, really you probably don't want one. 

David

Yup.




A bit lighter, though.  I got it a few years back from a silent key; looked last night at the bay of evil and was shocked at what people were asking for them.  Could have sworn I was seeing them for much less a few years back.  Yowsa!

-Pat

[Cubdriver]

Serial prefix 1246A, which would suggest a 1972 build date.

However, looking at the date codes on the reed relays in the input module indicates that it was built around 1976 so that may have been replaced at some stage. The serial number label on the module is still a 1246A prefix but it looks like it was peeled off and stuck down again.

It's quite possible that the input module was built in '76 but still had the old prefix - that only got updated for major engineering revisions, so if there were no changes it might stay the same for a long time.  Interesting that by '72 they were gold plating the boards.

-Pat

[Vince]

Yeah !!!

55 wired up, oscillates !

Put a couple 100K resistor and 1uF cap as an easy starting point with round numbers. Ran the math, gave a theoritical frequency of 5Hz, just the ball park I need so impeccable.

Works perfectly. ON time is spot on the OFF time as it ought to be, and frequency spot on 5Hz.  I am in heaven, it's a good start ! 

I calculated that a Frequency of about 20Hz (18.2Hz), would be good. Divided by 2^16 (x2 8 bit counter TTL chips cascaded), it would give me a one hour period.

So let's now fine tune this to get about 20Hz and add a pot to vary the duty cycle 

No RPN! But TI-85. Acceptable 

Actually really nice calculators and not the same heritage as the TI-83 as you can tell because the people who designed this didn't make the exponent key shifted. The screen was shit but they fixed that in the TI-86, which I had one of 

Ti 83 ? 86 ?! Don't know these...Hell I thought you were older than me not younger ! 
I was born in 1977, and AFAIR when I was at school the TI range was the TI 81, 82 and 85.

81 was a bit bottom of the barrel to my liking back then... well no honestly I just hated the colour. Much preferred the gray of the 82, so I convinced my mum to fork out 750 Francs for it. Was a huge sum for her, so I was very thankful. Well I think our applied physics teacher did also recommend the 82, to boot.
I loved it, it's well suited / tailored to engineering I find.  The 85 was so expensive that only one guy in my class had it. But I didn't like it because it was more suited to pure science and math... the K/B layout was not as practical for engineering I found. some basic functions were not as easy to access I found. Still, I was jealous of his 85... so 25 years later, recently, since they are now dirt cheap, I got one, got my revenge !    About 5 years ago.  When I received it, I had a "surprise"... inside the slide cover, there was something written  on it in big letters, by hand with I don't know, nail varnish or something similar, but silver looking. Like the conductive varnish to repair heated rear screens on cars.... anyway it read : " BASTARD !!! ". Yes in French it means the same as it does in English !

The seller omitted to inform of this...   
So I scrapped the varnish as best I could, but still !

Anyway as I said the 85 is a pain for basic engineering tasks I find, so I would much prefer using my own 82 of back in the day. Problem is that the LCD is kaput, vertical lines all over it 
I guess I could try to heat the cable see if that fixes it....


Anyway. Back at school, circa 1995 or 1996, other than that ONE class mate with a 85, there were two other mates that had a one-off as well. One had a Ti 92 ! A monster calculator, more of a portable computer than a calculator. huge screen, mega large K/B, needs to hands to operate it, a monster. Runs on a 68K CPU with 128K of RAM. Way too big and heavy in the bag to be practical, but boy in awe all we were, blown away by this thing  !!!! 

Then another guy had an HP 48 !  I remember I found it very "expensive". I mean the quality of the material and that of the build./construction, looked much better than the Ti. Also had a lovely comfy velvety pouch to keep it warm !    ISTR I found the quality of LCD better than the Ti. Had a cool Infra red link rather than a wired/cable for the Ti, however since the HP was so expensive he was the only one with a HP so had other calculator to connect to ! 
I do seem to remember however, that I didn't like the K/B at all ! It was 25 years ago so am not 100% sure, but I think I found the keys hard to press and noisy, compared to my Ti 82. Made typing both uncomfortable and slow. Again, old memories so could be off...

So yeah, at school it was either 82 for most of us, and 85 / HP 48 or Ti 92 for the more wealthy of us.

[mansaxel]

Nowt wrong with a TL072. Yeah I'm a fairly big fan of the TTi/Thurlby/Black Star etc etc stuff. It's not fancy, it's not beautiful but what it is roughly translates to very sound engineering. It's easy to maintain, well documented, cost effective and fit for purpose. Which is what you mostly want out of something.

The post was not meant to pick on the TL072 in any way; I quite like it; it's a cheap easy to handle jellybean IC; it works very well in most circuits, unless the circuit is bent towards something else. A typical alternative would be the NE5532, which also is a good opamp, but different in a few interesting ways. I feel there is a bit more solidity in line level buffers with the 5532, but it is subtle, and probably could be dealt with simply by tweaking the ancillary components. 

In stomp boxes, where people use harvested vintage JRC4552 and similar even cheaper  opamps "for their timbre", the TL072 is going to work just fine, as long as people don't know it's there. Issues  mostly arise because it's got a JFET input stage that borders on infinite impedance, so impedance taming needs doing a bit more aggressively. OTOH, that input, with its high impedance, is very good at capturing actual guitar tone; which of course might differ a fair bit from desired tone.

Opamps that are better of course exist, like DRV134 or the SSM2019 ones; but they are quite expensive and special in so many ways, and need some robust motivation to justify.

[Vince]

Fridge timer design project


OK so I added a pot to vary the duty cycle. Experimented a bit, tried a 22K, 47k and then a 100K.

Then played with the cap to target the 18.2Hz frequency to get about a period of an hour onc divided by two 8 bit counters.

So in the end I now have a 100K pot and the 1uF cap got replaced with a 100nF film cap  in  // with a 68nF in // with a 10nF.

No, 220nF is too big.

With these values, I can achieve the following :

- Duty cycle : target was 25 to 33%. I can get 24% to 40%, good.
- Frequency : target was 18.2. I get between 16.3Hz and 20.5Hz, so  pretty much my target +/- 10%, impeccable.


Very happy with that ! 

So :

1) Get an Oscillator working, square wave, 5V supply: CHECK

2) Variable Duty cycle : CHECK

3) Fine tuning to achieve target DC range and Frequency goals : CHECK


It's going well ! 

So now next, add a couple TTL 8 bit counters to get about an hour period.

[Cerebus]

PT100 & PT1000 sensors are thermistors;

That's not wrong in any sense, but they are usually called RTDs (Resistance Temperature Detectors) to differentiate them from NTC and PTC thermistors of mixed compounds.

[bd139]

Yeah !!!

55 wired up, oscillates !

Put a couple 100K resistor and 1uF cap as an easy starting point with round numbers. Ran the math, gave a theoritical frequency of 5Hz, just the ball park I need so impeccable.

Works perfectly. ON time is spot on the OFF time as it ought to be, and frequency spot on 5Hz.  I am in heaven, it's a good start ! 

I calculated that a Frequency of about 20Hz (18.2Hz), would be good. Divided by 2^16 (x2 8 bit counter TTL chips cascaded), it would give me a one hour period.

So let's now fine tune this to get about 20Hz and add a pot to vary the duty cycle 

No RPN! But TI-85. Acceptable 

Actually really nice calculators and not the same heritage as the TI-83 as you can tell because the people who designed this didn't make the exponent key shifted. The screen was shit but they fixed that in the TI-86, which I had one of 

Ti 83 ? 86 ?! Hell I though you were older than me not younger ! 
I was born in 1977, and AFAIR when I was at school the TI range was the TI 81, 82 and 85.

81 was a bit bottom of the barrel to my liking back then... well no honestly I just hated the colour. Much preferred the gray of the 82, so I convinced my mum to fork out 750 Francs for it. Was a huge sum for her, so I was very thankful. Well I think our applied physics teacher did also recommend the 82, to boot.
I loved it, it's well suited / tailored to engineering I find.  The 85 was so expensive that only one guy in my class had it. But I didn't like it because it was more suited to pure science and math... the K/B layout was not as practical for engineering I found. some basic functions were not as easy to access I found. Still, I was jealous of his 85... so 25 years later, recently, since they are now dirt cheap, I got one, got my revenge !    About 5 years ago.  When I received it, I had a "surprise"... inside the slide cover, there was something written  on it in big letters, by hand with I don't know, nail varnish or something similar, but silver looking. Like the conductive varnish to repair heated rear screens on cars.... anyway it read : " BASTARD !!! ". Yes in French it means the same as it does in English !

The seller omitted to inform of this...   
So I scrapped the varnish as best I could, but still !

Anyway as I said the 85 is a pain for basic engineering tasks I find, so I would much prefer using my own 82 of back in the day. Problem is that the LCD is kaput, vertical lines all over it 
I guess I could try to heat the cable see if that fixes it....


Anyway. Back at school, circa 1995 or 1996, other than that ONE class mate with a 85, there were two other mates that had a one-off as well. One had a Ti 92 ! A monster calculator, more of a portable computer than a calculator. huge screen, mega large K/B, needs to hands to operate it, a monster. Runs on a 68K CPU with 128K of RAM. Way too big and heavy in the bag to be practical, but boy in awe all we were, blown away by this thing  !!!! 

Then another guy had an HP 48 !  I remember I found it very "expensive". I mean the quality of the material and that of the build./construction, looked much better than the Ti. Also had a lovely comfy velvety pouch to keep it warm !    ISTR I found the quality of LCD better than the Ti. Had a cool Infra red link rather than a wired/cable for the Ti, however since the HP was so expensive he was the only one with a HP so had other calculator to connect to ! 
I do seem to remember however, that I didn't like the K/B at all ! It was 25 years ago so am not 100% sure, but I think I found the keys hard to press and noisy, compared to my Ti 82. Made typing both uncomfortable and slow. Again, old memories so could be off...

So yeah, at school it was either 82 for most of us, and 85 / HP 48 or Ti 92 for the more wealthy of us.

Slightly older than you but not much

To note, most of the calculators I have owned were well past the date they were introduced. This was because of a passing interest in collecting the infernal things. At one point I had 54 different calculators.

Also I used to buy and sell TI calculators here in medium quantities (tens of them at a time). There was a little niche where broke students where bouncing them for beer money at the end of the school or university year and then paying an absolute fortune for them just before it started because the entire supply was rinsed. I would at average buy for £7 and sell for £40. Do this 100 times and it's a good summer earner. This was a 6 week process of having stacks of mostly TI-83's on my desk, cleaning them up and refurbing them and testing them.

On the TI graphing line I have owned a TI-81, TI-82, more TI-83's than I can count, TI-85, TI-86, TI-89, TI-89 titanium, TI-92, Voyage 200, nSpire CAS.

But professionally and educationally speaking my timeline is roughly:

1984 -> Casio FX-82
1988 -> Casio FX-4000P
1990 -> TI-81
1993 -> HP 48G (this disappeared some time around 2003)
2003 to 2017 -> mostly excel intermission
2017 -> HP 50G / 35S (switch back and forth on and off)

The HP 35S is the nice sweet spot calculator now. It's programmable enough and I don't have to refer to the manual every bloody time I want to remember how to do something like the 50G 

As for the TI-92, I did indeed own one for a bit. Very nice machine but quite frankly it's still a TI-89 in a big box. That's not all bad but difficult to rationalise at the time and now. One of my university lecturers had one which was plugged into the large projector adapter which was quite fun to watch.

I actually wrote my first piece of commercial software on the TI-81. This was a horse racing gambling game which sold for 50p a copy and was typed in by hand from memory. If you paid me £2 extra you could get a set of PRNG seeds which guaranteed winners ... I had a real problem with software licensing when someone got a TI-82 and link cable

[Specmaster]

No RPN! But TI-85. Acceptable 

Actually really nice calculators and not the same heritage as the TI-83 as you can tell because the people who designed this didn't make the exponent key shifted. The screen was shit but they fixed that in the TI-86, which I had one of 

Ti 83 ? 86 ?! Don't know these...Hell I thought you were older than me not younger ! 
I was born in 1977, and AFAIR when I was at school the TI range was the TI 81, 82 and 85.

81 was a bit bottom of the barrel to my liking back then... well no honestly I just hated the colour. Much preferred the gray of the 82, so I convinced my mum to fork out 750 Francs for it. Was a huge sum for her, so I was very thankful. Well I think our applied physics teacher did also recommend the 82, to boot.
I loved it, it's well suited / tailored to engineering I find.  The 85 was so expensive that only one guy in my class had it. But I didn't like it because it was more suited to pure science and math... the K/B layout was not as practical for engineering I found. some basic functions were not as easy to access I found. Still, I was jealous of his 85... so 25 years later, recently, since they are now dirt cheap, I got one, got my revenge !    About 5 years ago.  When I received it, I had a "surprise"... inside the slide cover, there was something written  on it in big letters, by hand with I don't know, nail varnish or something similar, but silver looking. Like the conductive varnish to repair heated rear screens on cars.... anyway it read : " BASTARD !!! ". Yes in French it means the same as it does in English !

The seller omitted to inform of this...   
So I scrapped the varnish as best I could, but still !

Anyway as I said the 85 is a pain for basic engineering tasks I find, so I would much prefer using my own 82 of back in the day. Problem is that the LCD is kaput, vertical lines all over it 
I guess I could try to heat the cable see if that fixes it....


Anyway. Back at school, circa 1995 or 1996, other than that ONE class mate with a 85, there were two other mates that had a one-off as well. One had a Ti 92 ! A monster calculator, more of a portable computer than a calculator. huge screen, mega large K/B, needs to hands to operate it, a monster. Runs on a 68K CPU with 128K of RAM. Way too big and heavy in the bag to be practical, but boy in awe all we were, blown away by this thing  !!!! 

Then another guy had an HP 48 !  I remember I found it very "expensive". I mean the quality of the material and that of the build./construction, looked much better than the Ti. Also had a lovely comfy velvety pouch to keep it warm !    ISTR I found the quality of LCD better than the Ti. Had a cool Infra red link rather than a wired/cable for the Ti, however since the HP was so expensive he was the only one with a HP so had other calculator to connect to ! 
I do seem to remember however, that I didn't like the K/B at all ! It was 25 years ago so am not 100% sure, but I think I found the keys hard to press and noisy, compared to my Ti 82. Made typing both uncomfortable and slow. Again, old memories so could be off...

So yeah, at school it was either 82 for most of us, and 85 / HP 48 or Ti 92 for the more wealthy of us.

I have one of those TI92, terrible screen, black on greeny/grey does not work very well as far as contrast goes, making them hard to read, so it sits there in my collection and does not get used. Funny though because I always wanted one when they were new and expensive and out of my reach price wise, now I can afford one, it doesn't get used.

[mansaxel]

But professionally and educationally speaking my timeline is roughly:

1984 -> Casio FX-82
1988 -> Casio FX-4000P
1990 -> TI-81
1993 -> HP 48G (this disappeared some time around 2003)
2003 to 2017 -> mostly excel intermission
2017 -> HP 50G / 35S (switch back and forth on and off)

Mine is simpler:

1985 -> Sharp EL506P. Still have it, still works, still got original batteries in. Lacks enter key.
1990 to 2016 -> mostly pen and paper intermission; with some Excel, phone calc and awk thrown in
2020 -> SwissMicros DM42.

Oh, and Wife insists on TI calcs, so I got her a vintage TI-30 for either birthday or Christmas, don't remember.

Will probably get the children "hp" 35s calcs when they need something for school. Just to fuck with teacher brains.