Author Topic: T962 sudden temp failure  (Read 865 times)

0 Members and 1 Guest are viewing this topic.

Offline GrantIsCoolAllegedlyTopic starter

  • Newbie
  • Posts: 3
  • Country: gb
T962 sudden temp failure
« on: June 05, 2023, 02:19:26 pm »
Hi!

Long-time viewer of the channel and have come across the forums on my debugging travels over the years.

I bought one of those highly-regarded T-962 reflow ovens a few years ago, performed firmware, cold junction and kapton tape mods on it, and despite its shortcomings, it's more or less worked. Until now.

Temp suddenly isn't getting hot enough to reflow even Sn/Pb paste. Seems to be sensor-related. Wondering if you guys can suggest anything before I start throwing pennies at it.

•Think I can confirm, thanks to ohmmeter measurements and a temporary miswiring of the thermocouples (oops!), that triac, SSRs and quartz elements are all fine, as I managed to get the oven to be TOO hot, and blew up some caps. Good times. Back to too cold now.

•TCs both measure below 0.1mV at room temp, with 5-7 ohms across each. This is definitely wrong, no? Not an expert, but Google tells me it should be at least 0.8mV. My meter may be rubbish at such low voltages.

•Firmware reports a ~20 degrees difference between TCs (with Left being the clearly-wrong one).

•DS18B20 reads a sensible temperature.

•BOTH TCs failing seems unlikely to me, so I expect a common point of failure. Op-amp? Resistance measured in FB loop suggests op-amp gain on both channels SHOULD be around ~80, but one of the op-amp's output voltage doesn't reflect this. I'm never comfortable pointing straight at small-signal silicon as a point of failure! Would prefer to ask if op-amp failure is likely as I'll fork out for a better chip if that is the case, and £5+ is a lot of money in 2023 Britain. You could get at least half a biscuit for that.

•Can't really think of a particular event which may have caused it, besides setting a SAC305 profile to be slightly too high last week. Perhaps as a result of this failure creeping in, in hindsight.

Thanks for any guidance.
 

Offline mon2

  • Frequent Contributor
  • **
  • Posts: 460
  • Country: ca
Re: T962 sudden temp failure
« Reply #1 on: June 06, 2023, 12:22:09 am »
A few comments:

1) Review this website and document:

https://hackaday.com/2014/11/27/improving-the-t-962-reflow-oven/

2) Do not have this oven but we do own a Heller 1809EXL. Around 75 hours of use, behaved exactly like your unit. Room temp solder paste went in and came out at the end of the 9 zones in the same form. After some diagnostics, found the elements were ok but the thermocouple readings were not. Factory claimed schematics are available but then they said the board was outsourced (actually vendor is still around but ignored all of our requests for help) and that Heller no longer has the schematics. Attempted to reverse engineer the design and found that they used analog muxes and cycled the K-type temp sensors (16 of them) across the oven and sampled with a local A/D (something like that). After wasting a few days, asked the factory - how much and they stated $1500 USD. With some talk, 'a deal for us is $1200 USD'. Same unit from Shanghai (where it was built by shoddy employees) - $700 USD. We found the same controller in the surplus market in Texas @ $300 USD - bought 2 of them. Each worked immediately upon plug in.

Summary - our thermocouple controller board was measuring that the temp was being reached according to the profile, yet it was not.

Can you not just buy a spare controller board from Shenzhen? Suspecting it will not be too expensive. Can share some contacts in Shenzhen who may be able to supply this part. We use them for our spare parts to maintain our SMT line.
 

Offline GrantIsCoolAllegedlyTopic starter

  • Newbie
  • Posts: 3
  • Country: gb
Re: T962 sudden temp failure
« Reply #2 on: June 06, 2023, 01:27:05 pm »
Hi Mon,

Thanks for taking the time. The mods outlined in the Hackaday article have been performed, though the comments on that source and others have been quite helpful over time.

Buying a replacement controller board seems like a good option I haven't considered, though I hesitate as all the reflashing & mods would need to be performed again, only for a potential risk of a repeat incident.

I'm in a transition phase - volume of sales is such that hand-soldering is *just about* feasible, but it's getting a bit impractically time-consuming during busy periods. So I'm not at such a loss that I'd be in serious trouble if this isn't resolved quickly, but it does impede expansion. Even the surplus market price for your controller board exceeds the cost of the T-962 itself, whose overall cheapness makes me pessimistic about the likelihood of finding a separate controller. I'll have a look, though.

Assembly quality is extremely poor, with resistors doing gymnastics and whatnot.

Just thinking:
"Specified Temperature Range [for TTLC27L2]:
0°C to 70°C . . . 3 V to 16 V"

That's no good.

Cheers! Food for thought.

 

Offline GrantIsCoolAllegedlyTopic starter

  • Newbie
  • Posts: 3
  • Country: gb
Re: T962 sudden temp failure
« Reply #3 on: June 14, 2023, 03:11:21 pm »
Haven't tried a proper reflow yet, but replacing the TLC27L2 with an AD8552ARZ has fixed the erroneous reading on one channel (now both reading room temperature when first powered on) and IR elements are turning on & off much more in-line with expected behaviour upon a dry run. Fingers crossed.

TLC27L2's datasheet: "Input Offset Voltage Drift . . . Typically 0.1 μV/Month." That'll do it...

AD8552's datasheet: "The AD8552 provides the benefits previously found only in expensive auto-zeroing or chopper-stabilized amplifiers. Using Analog Devices, Inc. topology, these new zero-drift amplifiers combine low cost with high accuracy." Score!
« Last Edit: June 14, 2023, 03:32:04 pm by GrantIsCoolAllegedly »
 
The following users thanked this post: Kean


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf