Rancilio Silvia V3 - Quick and dirty ~$5 & 2hr full internal insulation

[nikko.the.scorpio]

Hi,

Firstly, I don't intend this to be a thread discussing the merits of insulating a simple single boiler machine like the Silvia - rather just to point out how quick and cheap it is to fully insulate all the appropriate internals IF that takes your fancy.

I know insulating PID equipped Silvia's is quite common - however I did mine on the basis of keeping more of the generated heat where I want it (in the brass/water etc) - while it may seem simple this is where the heat is wanted and should mean less 'brew cycles' in my warm up time etc. I know there is the argument about thermal overshoot and insulation making this more extreme BUT I tend to do a 2-3 sec cooling flush prior to extraction if you waited a given period of time you'd have to extend this to adjust for the improved heat retention. Anyway as I said I didn't want to do a pro vs cons as suffice to say I felt it made sense for me.

Thanks to Greg Wormald (hope I spelt that right) - as several posts of his keyed me onto Insul-Bright as a suitable material for internal insulation of espresso machines. It's essentially polyester wadding with an aluminium/mylar layer interfaced within it. Many posts discussing it's heat rating etc - I got mine at Spotlight ~$16/m for the ~115cm wide roll they sell (they also sell a ~58cmm wide roll but the larger one is better value).

I bought a metre of it - though thats WAAAAAAAAAAAAAAY more than required - my wife will make many oven gloves with the remainder.

Now FWIW I did cheat a little - though it really wasn't that handy in the end. I downloaded the image from PID Silvia of their pre-cut V3 insulation. If you enlarge this 190% and print it up it's about bang on the right size for the require pieces. This is handy for the top of the boiler but I did find that for the rest of the Silva I just cut bits off as required - BUT if thats your thing 190% enlargement is right to allow you to make almost perfect templates.

Fully dismantled the Silvia (when you remove the cables etc make sure you number them or similar so you know what goes where). By measuring and using the piece of steel that goes over the grouphead section I made the right shape for their, did right around the outside of the group with a single piece tripled over, even tucked a bunch of small bits around behind it. You have to play around a little so you don't get lil bits coming out down the bottom - but it's pretty easy to do. The grouphead was very well packed - having that piece of steel over the top really allows you to load it up and ensure it's very well insulated.

The boiler was next. Measured it's height and circumference, cut a single piece and went round twice. Cut lil slits/sections out as I went around to allow me to get around the two lil bits that protude from the side. Connected two cable ties to each other to allow me to get around the outside and hold it in place. Insul-Bright is actually VERY friendly to work with and due to the mylar/foil it will be able to be pushed/molded into place to a certain degree. I made sure the sides lengths were a few cm more than required so I could put it down and cover the full length of the boiler.

Then used the templates to do the top - this was handy. Important to cut the 2 small holes in the right spot as this really holds it in place. Triple layer up top.

The steam arm was easy, used a very long inch thick strip and wrapped over and over. Did a double layer on this - and it was essentially already doubled up from the wind over technique so this is very well insulated. Even did the ball joint section with a nice wrap around - again very roughly cut out a longish inch wide strip and you just wind it around like you might a bandage etc.

The boiler connection/OPV was last but again inch wide strip and round this a bunch of times and couldn't be easier!

Then wiped out the exces polyester fluff that had come off - also vacuumed it out a lil. Screwed everything back into place and carefully reconnected all the electricals, making sure not to accidently catch the mylar foil into a connection (this was my only concern).

All cleaned up pretty easily and no polyester fluff showing at all. My one slight concern was whether there'd accidently be any shortcircuits by the mylar foil around the terminals at the top of the boiler but this was thankfully not an issue - powered up fine!

IMG_0103 by

It's no beauty queen BUT I estimate the cost of the material used was around $2-3. It took me around 2hrs - but thats me fiddle farting with the templates etc and you really don't actually need those. I won't use it till tomorrow morning but I anticipate it should be a tiny bit more temperature stable and also require less pre-heating - perhaps marginal benefits but it was a small amount of work and other users have reported using Insul-Bright for years with no issues.

There was no sewing, glue or anything other than a few cable ties. So whilst I don't doubt the closed cell silicone is a superior material - it's cost + inavailibility are negatives for it. I was also able to use the Insul-Bright in a few spots the silicon couldn't get (like behind the grouphead/boiler slot) so I don't think it'd be that worse off in overall performance.

Anyway see what you think but it's a very easy overall project that anyone of entry-average skill could easily do in an afternoon.

Cheers, Nick

PS. As stated earlier full credit to Greg for the idea.

[nikko.the.scorpio]

Used this morning - noticeable how much longer between heating cycles it is now. Not exactly sure but I'd estimate a significant difference.

Also when I pull the initial cooling flush (prior to an extraction) there's a tad more steam out via the group and I also thought the steam arm pressure was higher than normal - it definitely had a lot less water built up in the arm when I purged it several times prior to steaming the milk. All things that would seem logical from improved thermal retention.

[Dimal]

Is your Silvia PID Controlled "nts"?

If not, you're probably going to find that the brew water temperature is going to be more elevated than is ideal. That's why it's popular to do this with PID Controlled units, brew water temp. is much more easily controlled and maintained...

Mal.

[nikko.the.scorpio]

Is your Silvia PID Controlled "nts"?

If not, you're probably going to find that the brew water temperature is going to be more elevated than is ideal. That's why it's popular to do this with PID Controlled units, brew water temp. is much more easily controlled and maintained...

Mal.

Hi Mal,

No PID on mine (possible option down the track if I can figure out how to install one of the ~$USD30 kits). Understand exactly what you're saying however I'd have thought that the temperature overshoot will still be the same - the improved thermal retention will just mean that it's slower to get down to a certain temp after this. So for example if I use the temp figures posted here by another user:
Silvia Mega Mod by Avi. Rancilio Silvia coffee machine modding PID Silvia, Brew and Steam control

Heating:
Heating element lamp turns ON @ 94.0°C (201.2°F).
Heating element lamp turns OFF @ 109.4°C (228.9°F).
After it turns off, overshoot temperature reached 121.9°C (251.4°F) before it starts dropping.

My boiler's numbers might be a bit different but you'd expect them to be in the same ballpark. So the simple boiler control on the Silvia will still turn off at the same set temperature (as I'm assuming it's temp controlled and not a time controlled heat up) - hence the off temp and overshoot should still be essentially the same - however the difference comes in where an uninsulated system might reduce in temp at say 20 degrees/minute mine might now only reduce at 10 degrees/min.

I might be wrong but it shouldn't get any hotter etc - just that if one was temp surfing by using time after the boiler cycles off then you'd have to wait considerably longer.

However as stated I prefer to do a 3-4sec cooling flush immediately after the boiler goes off and then extract straight away. I'd have thought this approach would not have changed the actual water temp at all - but could be wrong.

What I might have to do is test the brew head exit temperature using this methodology and see if it hits the apparently recommended temp of 88c. I'll check and see how that is - thank you for prompting me on this. :-)

[papagoose]

Well the point of the PID is that there wouldn't be a temperature overshoot. With a set point of say 106C, power to the element will be pulsed on the order of seconds, so the temperature will actually just sit constant at 106. Where you could still see an overshoot is when you first turn the machine on, if the PID constants are tuned to an uninsulated boiler. But as I understand, the PID 'learns' the properties of the boiler and will compensate for this, on the order of minutes. So with a PID, you might find that with an uninsulated boiler, it's on for 0.1 seconds, off for 0.5 seconds, but with an insulated boiler, it's on for 0.1 seconds, off for 0.9 seconds. But in both cases, the boiler will sit at the set point, to within the precision of your thermocouple. The key difference then is that unlike the stock thermostat, a PID is able to adapt to the difference in heat retention.

[nikko.the.scorpio]

Well the point of the PID is that there wouldn't be a temperature overshoot. With a set point of say 106C, power to the element will be pulsed on the order of seconds, so the temperature will actually just sit constant at 106. Where you could still see an overshoot is when you first turn the machine on, if the PID constants are tuned to an uninsulated boiler. But as I understand, the PID 'learns' the properties of the boiler and will compensate for this, on the order of minutes. So with a PID, you might find that with an uninsulated boiler, it's on for 0.1 seconds, off for 0.5 seconds, but with an insulated boiler, it's on for 0.1 seconds, off for 0.9 seconds. But in both cases, the boiler will sit at the set point, to within the precision of your thermocouple. The key difference then is that unlike the stock thermostat, a PID is able to adapt to the difference in heat retention.

With all due respect I've been aware for sometime of what the purpose of a PID is and what qualities it brings to a simple thermostat controlled machine like the Silvia.

I've simply not fitted one as the kits are (IMHO) cost prohibitive vs the actual value ($400-500) of a used Silvia. Now I'm unsure if the cheaper China Ebay kits that sell for $30-50 are even vaguely comparable but if they are then perhaps I'll consider one down the track. Till then temp surfing isn't that hard and I can't see any reason why the 3-4 sec flush won't bring the normal overshoot down to the right point.

The insulation was done to essentially bring the overall required warmup time down (which I'm certain it's done) and reduce the number of boiler cycles needed - amongst other benefits. Simple job, simple benefits. Like i said I'll test the actual temp of water out of the group to make sure my methodology is working. :-)

[nikko.the.scorpio]

Ok so tested this morning after I'd made our morning coffees. The machine had plenty of time to warm up (had previously been on for ~45mins) and I'd also allowed another 20mins or so to cool back down from steaming (as that's the last thing I do in my workflow).

Alas worst fears concerned as even after a solid 5 second plus cooling flush it was still showing water temp through the grouphead of around 92-93c!!! Now my understanding is that it is MEANT to be 88c (apparently thats what some Italian Espresso body advises is the best temp). Temps were taken using a K-type thermocouple into a preheated small plastic vessel, so very accurate and instant.

So............................................ :-/

Haha well on a positive note I suppose it's safe to say that the insulation on the boiler is working very well! But seriously now I have to kind of figure out how best to get this down to the 88c needed for best extractions. Is a bit of a bummer that I didn't note the same temps etc PRIOR to putting the insulation on.......as it might just be running high anyway...but lets assume thats not the case.

To me there's really only 3 or so viable options:

1) Remove the insulation from ONLY the boiler: the grouphead, steam arm, OPV insulation can stay as that more keeps existing heat in - but the insulation on the top & sides of the boiler is causing whats already a machine design thats KNOWN to overshoot it's target optimum heat mark by 10c+ to perhaps go even further above this. Haha I've already been wrong once BUT I'd assume that taking it off the boiler SHOULD bring it back to a stock range of heat on the normal boiler cycle.

2) Fit a PID & leave insulation on: A PID for the Silvia is a good idea - thats why so many do it - but my sole hang up with it has always been that for the PID kits e.g Watlow etc they're 50%+ of the market value of the machine. And to me thats illogical (better off selling and buying a Lelit etc with a factory PID). HOWEVER it would be viable if I could get through all the double dutch on which 'Fleabay' bits and pieces are acceptable to buy and fit. I'm pretty reasonable with fiddly installs etc so no fear in that regard. Seems that all the pieces needed should be available for under $50 - but obviously you have a few hours fitting and putting the right settings on etc. This would be the best overall outcome but I don't know if it's actually viable.

3) Completely redo my extraction timing: Essentially what I mean by this is instead of brewing based on the boiler light going off, then applying a cooling flush for 3sec then extracting I'd have to basically brew on the bottom of the heating curve. e.g waiting till the light goes ON then waiting perhaps 5 sec THEN extracting. Thats a bit of an out of the box idea BUT I was checking the temps when the brew cycle light went back on and they were around 86-87c. So thats a LOT closer to the 88c needed than at the top of the cycle.

Hmmm anyway not really sure what to do right now....both 1 & 3 would be very simple to do - though I have to say I'd feel a bit of a sellout going with #1 as the logical part of me thinks there has to be a more effective way to find the right spot in the cycle to extract. #3 just requires a bit of testing and 'rebooting' my current methods but thats pretty simple to do.

Personally I'd prefer to fit the PID but thought I've found a few forum posts on fitting the cheaper PID kits/pieces I've not had much luck getting anything recent etc and it seems like a tricky matter unless you're 100% sure you're getting correct advice. Haha so very open to feedback if anyone knows a good article or has recently fitted something like that. :-)

Thanks again for the great advice, especially Dimal as you were again correct and I really appreciate you drawing this to my attention. :-)

[saeco_user]

I bought an elcheapo MYPIN Dual Digital PID Temperature Controller + SSR + K Sensor Thermocouple Probe TE75 from evilbay ($50) last Dec. and installed it a few weeks ago in my Quaha.
It certainly stabilised the boiler temperature dramatically prior to pulling a shot.

The sensor had a metal braid sheath and was mounted inside a screw in terminal. I removed these and mounted the naked sensor on the boiler.

It was very simple to install / test - at least on a temporary basis. I haven't worked out yet the best way of installing it permanently, there is less room inside the Quaha than the Silvia, and I haven't found a decent looking box to mount it in.

The only small concern with the MYPIN PID is that the temperature measured with a Thermocouple seems to drift as the PID warms up compared to a reference thermocouple connected to a fluke thermometer.
This is only by a couple of degrees so not particularly serious. At some stage I will try an RTD sensor to see if the drift it is due to a lack of proper thermocouple compensation in the PID.

I recently removed the PID after cleaning and resealing all the machine internals (PID and SSR were housed in a large pink lunchbox sitting on top of the machine with separate mains lead and sensor/SSR output lead coming out of the box - very ugly). Having to go back to temperature surfing is now a pain, wastes a fair bit of water, and I miss knowing at a glance what the boiler temperature is. I am very keen to re-PID this weekend, hopefully in a more permanent fashion.

Taan

[nikko.the.scorpio]

Taan,

Great reply and much thanks. REALLY appreciate you taking the time to pop this up.

I might try and either 'resurrect' an old thread on the cheaper PID kits or get a new oone up - as I think it's an area that appeals to a lot of folks or atleast would if they knew they could get a kit so cost effectively and saw that others put them in without too much messing around.

Would really love to find out more about your install etc o do feel free to post up if/when you do it as I for one would definitely be keen to see. Much thanks, Nick

[nikko.the.scorpio]

Hmmm just doing a bit of research on PID kits and stumbled across this video by PIDSilvia (who's probably the best known modifier of Silvia's) - what I noticed was that he has a temp display for the water coming through his grouphead. Now during both of the extractions this is showing basically 200F or 93c. Which is basically exactly what I was getting (and subsequently freaking out about) when measured this morning after a quick flush.

https://www.youtube.com/watch?v=D5xtWEXOT58

So........perhaps it's not as dire as I'd thought as look I've got to assume he knows his stuff and that must be a pretty decent temp to be using - several assumptions there BUT I think they're relatively safe! Probably a lot safer than saying as I wasn't getting 88c that I need to fit a PID kit (which still might be a decent idea but not as desperate as I'd feared!).

FWIW I've tossed a general thread up on any DIY PID kits - so would greatly welcome any thoughts or discussion on those as I really think a lot of other users would be interested in reading about those so I'd urge discussion on them placed there so others can also benefit. :-)
http://coffeesnobs.com.au/brewing-eq...-new-post.html

[Dimal]

G'day Nick...

Brew Water Temp. of 92-93 Deg. is close to perfect mate so nothing to worry about there, depending on how you are measuring that of course...

Mal.

[Henriklam]

Good work Nick! This will be on my schedule for the coming months