I was looking around at PID kits and it seems like they are extremely expensive compared to the sum of individual component costs. A PID kit can cost somewhere around 200-300$ and often does not include installation.

A dual temperature PID controller starts at less than 50$ shipped on Ebay (http://shop.ebay.com/i.html?_nkw=dual+p...m270.l1313). A relay will probably cost less than 5$ (http://www.radioshack.com/search/index....relay&sr=1). Wires, terminals, temperature probes cannot possibly add up to more than 50$. Extra cosmetic accessories like Velcro, wire tubing/zip ties will add up to around 10 bucks. Total cost =~ 120$.

Why is there such a huge markup? Is it because the PID kits offers a higher end PID controller? Are people are hesitant to buy their own parts? Originally I thought it was because the average person lacks the technical capabilities to do the installation, but the kits often require self-installation. What do you guys think?

From a technical perspective, I'm not sure I understand the difference between a more expensive PID and a cheaper one. I did some brief PID research (including a discussion about PIDs with an individual with a EE Masters degree focusing on Controllers) for a DIY Sous Vide project and it was to my understanding that for the most part there are industry standards for PID algorithms. Why the cost difference?

As for accuracy - most(?) PID kits seem to use type-K thermocouples. While thermocouples have a faster response rate than resistance temperature detectors (RTDs), they are not known for accuracy down to a single degree, and often have a +-2 degree accuracy. In addition to this, there is temperature drift involved with thermocouples, sometimes after hours of use. Considering that espresso brewing is affected by as little as a 1 degree change, this seems a little sloppy to me.

With that in mind, are PID temperature probes normally mounted outside the boiler? I believe this is the case for PID kits, but what about for the PIDs within the higher-end professional machines? If the probes are mounted outside the boiler/grouphead, then the actual brewing temperature will be time-offset from the PID temperature reading because heat needs time to transfer via conduction. When the heating element is on, it will take a bit of time to heat the water, then the water takes a bit of time to change the temperature of the grouphead/boiler surface. As a result, a significant time lag can be seen causing errors in the process.

Last but not least - has anyone tried to use PIDs to control the pump pressure? It seems like this could be done with the help of a digital pressure gauge and a motor/servo. This might be a little too far fetched of an idea though.


Disclaimer: I am not an expert in espresso machines, PIDs, temperature probes, or hardware design. I have done a lot of reading on these topics and merely understand the concepts, so maybe the things I suggested are completely ridiculous and/or stupid and someone will come and correct me regarding my misconceptions.

PID controllers and SSRs are manufactured by the 100,000s in China; PID kits for espresso machines are hand made one at a time, for the breathtakingly large market of a few thousand potential customers. In other words, they'd be expensive even of the controller came down to a few bucks. If you want to save the money, you're welcome to install your own PID.

Just a few quick thoughts on your questions:

1) You get what you pay for with quality of PIDs with respect to longevity, ease of use, button quality (I had a cheap PID where the jelly buttons all broke within a year). PIDs are just like anything else, the more you pay the better quality you get (usually).

2a) Sure a k type thermocouple may only accurate to +-1 degree, but that doesn't mean its precision is +-1 degree, and since taste is determining factor for temp setting ("lets try it up a half degree" rather than "this espresso HAS to be brewed at 101.4 degrees"), it is precision that matters more.

2b) At the temps we are measuring, drift is not nearly as extreme for a high quality thermocouple... also, the higher quality PIDs have a well designed circuit to compensate for drift (see answer 1).

3) Kits for home machines use external sensors b/c boilers are harder to gain access to wrt interior water temp... thermal offset is consistent though, so its not that huge of a deal. The time offset... well, the thermal system can't shift too much if you are keeping the physical boiler 'container' temp consistent since the heat transfer in such a small boiler is fairly rapid (many boiler 'containers' actually contain the heating elements themselves), a PID takes care of this offset issue b/c it 'learns' the properties of the system (a good PID does). PIDs on commercial machines and prosumer machines usually always directly measure water temp in the boiler.

4) Re pump pressure PID... (edit: removed poor/inaccurate communication) If you will browse and search the forums, the proper (and the very challenging ways of pressure profiling) ways to do this have been discussed ad nauseam... many really brilliant controls people here.

PS: Thanks for the disclaimer, the intertubes don't create experts, they just fool people into thinking they are.

Thanks Karl. I believe you answered most of my questions. It's comforting to know that my concerns were mostly theoretical, and that in practice they don't matter as much as it seems like they would.

Also, I did a quick search on controlling pump pressure as well as some reading on "pressure profiles" and now understand that a PID setup for pump pressure will be fairly useless because the flow rate is not linearly correlated with pump pressure. As a result of this, only simpler voltage mods to control pressure are worthwhile.

Hooray for learning!

Quite a few of us have sourced the parts and fitted a pid to an espresso machine. For a while the Watlow 935a pid used could be had at a good price but the source seems to have dried up. I pid'ed one sb/du machine for ~$50.

If you want a slice of cake but have neither the ability or the desire to bake one from scratch then have the bakery make one for you. You will be assured of having a finished eatable cake but it will cost a bit more. However if wish to cook for yourself have at it and save a buck at the same time.

My experience is with two PIDs: two used Watlow 935a and a new Delta DTB PID I bought from Factorymation. The Delta unit is a cheapish Chinese unit.

I was not impressed with the Delta's software, particularly the ramp/soak programming part. It regularly sends corrupted data to the controller, necessitating a reboot of the controller and restart of the software. The manual is not as comprehensive as the Watlow, and it also cannot accept times below 1 minute (ramp/soak) or 1 second (Derivative time parameter) like the Watlow can.

When I PIDed my SBDU machine http://coffeegeek.com/forums/espresso/machinemods/483786, I looked for a used Watlow rather than a new cheapo unit. I think it delivers better value for the money.

I have a Love 32B on the Vibiemme HX. If it's not identical to the similar Delta, it's as close as damnitall is to swearing.

It's been flawless for 4 years.

danaleighton wrote:It regularly sends corrupted data to the controller, necessitating a reboot of the controller and restart of the software.

Sounds like the device is sending data to an external computer? It's been my experience that bundled applications with inexpensive hardware are pretty useless at error handling. There's a reason Flukes cost more.

It's about matching the control to the application.

Ktzero3 wrote:I did a quick search on controlling pump pressure as well as some reading on "pressure profiles" and now understand that a PID setup for pump pressure will be fairly useless because the flow rate is not linearly correlated with pump pressure. As a result of this, only simpler voltage mods to control pressure are worthwhile.

No idea what you're talking about. Several of us have had success with PID control of pump pressure. The fact that flow rate is not "linearly correlated with pump pressure" is true but besides the point.

Sorry AndyS, I think I mislead him with that... my bad. After reviewing his question, I think I was at the time answering 'pressure profiling' rather than 'pressure precision' which was inaccurate of me. The initial road that I sent him down was not his fault, but mine. Sorry.

Another possible difference between pro-quality and eBay specials: I remember reading somewhere that cheap controllers might not like having the measurement junction of the thermocouple in contact with grounded metal. To minimise the response time of the system, as you mentioned, it's best to use an unsheathed t/c directly in contact with the (grounded) boiler, so that could be a problem. It would be interesting to know if anyone's come across that problem here, though a brief search didn't throw it up...