Heated Digital Stirplates

[Mardoo]

When your wife asks you why you've been watching your Stirplate for an hour, just say, "Farmin," in your best Clint Eastwood voice. Because you will.

 

[durham70]

Picked mine up from the PO tonight Got yeast but no flask or DME yet

 

[Frothy1]

Im pretty impressed.

Also I dont have to squeeze it into the ferment fridge.

Just have to sort my temp adjustment as I didn't expect it to go absolutely freakin gang busters.

 

[peteru]

Got my Yeast Forge on Friday. Delivered about 22 hours after ordering - I like buying from Australia.

I didn't get to play until Saturday afternoon, but here are some first impressions:

* The retail display packaging, even though simple, is a nice touch that makes a good first impression. It's just a bit of cardboard stapled to a plastic bag, but it certainly has a classier feel than if the product was just in bubble wrap.

* I like the simple clean lines of the enclosure. If at all possible, it would look nicer if the stress relief grommet for the probe cable was less bulky. It's probably oversized, since the probe cable can still slide in and out even though the nut is tightened all the way. On the other hand, reliability and durability is preferable, so if alternatives would compromise the product quality, stick with the current option.

* The enclosure needs some non-slip feet / strips attached to the bottom. Pressing the buttons requires enough pressure to cause the Yeast Forge to slide all over the place on most surfaces. This is easy enough to retrofit at home with self adhesive rubber feet, but it would be nicer if it was a standard feature.

* First thing I tried was to run the Yeast Forge without a load. I was surprised at how noisy it was. The sound is reminiscent of a computer fan rattling due to worn out bearings. I'm not sure if that is normal due to the magnets attached to the fan, causing it to be slightly out of balance. It spins on all speeds from 200 to 2,000 RPM, with and without load, so the noise does not seem to indicate a complete failure. The noise is not a horrible squealing or screeching noise that a completely buggered fan would make, more like the first signs of trouble on a fan that has been in service for a long time. It would be interesting to hear from other users whether their Yeast Forge is silent or if it makes noise and how much.

* I think the LED display refresh algorithm may have a bug. I set the temperature to 20C and speed to 2,000 RPM. I then observe the LEDs. What I see is that all the LEDs flicker off on a semi-regular basis approximately every 1/2 - 3/4 of a second. At a guess, I would say that all the LEDs are turned off and then back on again very rapidly. It is so quick that some people may perceive this as a slight dimming in LED brightness every now and then. That could probably be fixed with a firmware upgrade.

* The temperature probe calibration seems to be 2C off. I set the temperature to 20C and attached the probe to the vessel using rubber band and bubble wrap as suggested. I've actually got three probes in there, right next to each other. The other two are reading 22C, even though Yeast Forge has been set to 20C. Ambient temperature has gone from 18C down to 15C over the last few hours. The higher temperature does not appear to be due to yeast activity because the heating light comes on every now and then.

All-in-all, I like the new toy. The above are suggestions for tweaks so that the product can improve in the future, but it's a pretty good thing to have in the current incarnation.

 

[Beamer]

Will be placing an order Tuesday next week, I love a tax return, even better when SWMBO is the accountant.

 

[megabyte]

Got my Yeast Forge on Friday. Delivered about 22 hours after ordering - I like buying from Australia.

I didn't get to play until Saturday afternoon, but here are some first impressions:

* The retail display packaging, even though simple, is a nice touch that makes a good first impression. It's just a bit of cardboard stapled to a plastic bag, but it certainly has a classier feel than if the product was just in bubble wrap.

* I like the simple clean lines of the enclosure. If at all possible, it would look nicer if the stress relief grommet for the probe cable was less bulky. It's probably oversized, since the probe cable can still slide in and out even though the nut is tightened all the way. On the other hand, reliability and durability is preferable, so if alternatives would compromise the product quality, stick with the current option.

* The enclosure needs some non-slip feet / strips attached to the bottom. Pressing the buttons requires enough pressure to cause the Yeast Forge to slide all over the place on most surfaces. This is easy enough to retrofit at home with self adhesive rubber feet, but it would be nicer if it was a standard feature.

* First thing I tried was to run the Yeast Forge without a load. I was surprised at how noisy it was. The sound is reminiscent of a computer fan rattling due to worn out bearings. I'm not sure if that is normal due to the magnets attached to the fan, causing it to be slightly out of balance. It spins on all speeds from 200 to 2,000 RPM, with and without load, so the noise does not seem to indicate a complete failure. The noise is not a horrible squealing or screeching noise that a completely buggered fan would make, more like the first signs of trouble on a fan that has been in service for a long time. It would be interesting to hear from other users whether their Yeast Forge is silent or if it makes noise and how much.

* I think the LED display refresh algorithm may have a bug. I set the temperature to 20C and speed to 2,000 RPM. I then observe the LEDs. What I see is that all the LEDs flicker off on a semi-regular basis approximately every 1/2 - 3/4 of a second. At a guess, I would say that all the LEDs are turned off and then back on again very rapidly. It is so quick that some people may perceive this as a slight dimming in LED brightness every now and then. That could probably be fixed with a firmware upgrade.

* The temperature probe calibration seems to be 2C off. I set the temperature to 20C and attached the probe to the vessel using rubber band and bubble wrap as suggested. I've actually got three probes in there, right next to each other. The other two are reading 22C, even though Yeast Forge has been set to 20C. Ambient temperature has gone from 18C down to 15C over the last few hours. The higher temperature does not appear to be due to yeast activity because the heating light comes on every now and then.

All-in-all, I like the new toy. The above are suggestions for tweaks so that the product can improve in the future, but it's a pretty good thing to have in the current incarnation.

Hi Peteru, thanks for the feedback about the Yeast forge. It's great to hear your likes and dislikes so we can make improvements in the future.

For the grommet, I agree right it looks a bit big. Thanks for bringing this up, sometimes you just need to hear someone else's perspective to make up your mind. We've used a PG-7 which is the smallest we can source, but in future it will be better if we can find an alternative cable gland that's more suited to the small temperature probe cable. I'll look into it! Non-slip feet would be nice too. Unfortunately this enclosure doesn't have any recesses they fit into, but we could have them machined in future so it's a great suggestion.

For the sound, there's normally a quiet sort of ticking noise which comes from the motor windings as the PWM algorithm tries to keep it running at a constant speed. We use the fan's blades to determine the fan's speed and they're never entirely symmetrical so the controller is constantly varying the motor's power which you can hear especially at low RPM. Don't be concerned about longevity though, we've been running tests for a couple of years with a demo unit spinning up and down, day and night and we're yet to burn out the fan or damage the bearings. Of course there's a lifetime warranty too.

The LEDs do flicker a bit. It's actually when the external probe takes its temperature measurement. It uses a DS18B20 sensor which is timing critical so we have to pause the display for a few milliseconds each time we take a temperature reading and you can see that as a small flicker every 3/4 of a second or so. In the future we could try and talk to the probe by hacking the USART port but we'll need a new PCB design for that. Alternatively we could use a thermistor or analog temperature probe, but I really like that the DS18B20 is a known quantity and has great precision and accuracy.

Regarding the calibration, the temperature probe's spec is +/- 0.5˚C so the more you can insulate it from ambient the better. I think this is one of the main points where we can improve our design, either using a non-contact temperature sensor or supplying some kind of jacket in the future.

 

[bradsbrew]

Anyone else have problems changing the temp? Following the instructions, i have turned the spinning to off then hold down the button but i only get 2 beeps and then nothing, the whole thing shuts down and i have to turn off and remove the power cord then turn it back on to get it spinning again but still can't get to temp change program.

 

[Frothy1]

Anyone else have problems changing the temp? Following the instructions, i have turned the spinning to off then hold down the button but i only get 2 beeps and then nothing, the whole thing shuts down and i have to turn off and remove the power cord then turn it back on to get it spinning again but still can't get to temp change program.

Yep, Its happened a couple of times.

Give it 10 seconds or so without power seems to do the trick.

 

[bradsbrew]

Yep, Its happened a couple of times.

Give it 10 seconds or so without power seems to do the trick.

Had it powered off for a few minutes and cannot get into the temp setting at all now.

 

[megabyte]

Try this.. Set speed to maximum, then press up again, then bring it back down back to zero and all will be well again.

If it's what I think it is, the problem arises when the power is interrupted at a time that the yeast forge is storing its speed and the voltage is too low for the chip to complete the write. We've recently updated the programming of the chips to enable "Brown Out Detection" and this appears to have solved it completely. Can you please PM me your purchase details and I'll know for sure whether it's pre-update. We'll sort you out!

Frothy1 same goes, please try this too if it happens again!

 

[bradsbrew]

Try this.. Set speed to maximum, then press up again, then bring it back down back to zero and all will be well again.

If it's what I think it is, the problem arises when the power is interrupted at a time that the yeast forge is storing its speed and the voltage is too low for the chip to complete the write. We've recently updated the programming of the chips to enable "Brown Out Detection" and this appears to have solved it completely. Can you please PM me your purchase details and I'll know for sure whether it's pre-update. We'll sort you out!

Frothy1 same goes, please try this too if it happens again!

Just tried this and still have the same problem. Only just purchased on Thursday as a pre order and received Friday, so the recent upgrade has not helped this issue. To be totally honest at a price point of $180, i did not expect to be a beta tester!

 

[megabyte]

Drat. I understand your frustration and I'll do everything I can to sort this out for you ASAP.

I'll PM you to get your order details and hopefully we can get to the bottom of this very quickly.

 

[megabyte]

OK I think everyone's in bed now but I've found the problem.

The issue occurs when holding the temperature set (down) button longer than necessary (about an extra second after the two beeps). The Yeast Forge goes into failsafe and you have to cycle the power to reset it.

It was a rookie programming mistake and I'm in contact with bradsbrew and Frothy1 hoping to organise a firmware update for them. I'll track down recent orders tomorrow (today). If others are experiencing this problem, please PM me your order details and it will be my pleasure to sort this out for you.

I am humbled by my mistake, but ever grateful for the timely and constructive feedback provided.

 

[peteru]

thanks for the feedback about the Yeast forge. It's great to hear your likes and dislikes so we can make improvements in the future.

Thanks for the detailed responses. I'm hoping that the feedback will contribute to your success and you can improve existing products and make enough money to keep producing more cool things.

For the sound, there's normally a quiet sort of ticking noise which comes from the motor windings as the PWM algorithm tries to keep it running at a constant speed. We use the fan's blades to determine the fan's speed and they're never entirely symmetrical so the controller is constantly varying the motor's power which you can hear especially at low RPM.

I think your description of the source of the sound is matching up with what I am hearing. So with the origin of the noise explained, I get to my usual bit where I say, "Geez, I wish I could get access to the firmware source code and start experimenting with the algorithms." I peeked inside and it looks like the six pin probe connector doubles up as an SPI in-circuit programming connector. I'd probably need to get a suitable programming cable, which at a guess would be a 5V interface and some kind of ATtiny88 development environment.

I'm thinking that there might be a slight mismatch between the usage pattern and the implementation details. It sounds like the implementation looks at the RPM of the fan and attempts to adjust the speed by varying the PWM ratio continuously to stay within a certain tolerance of the target speed. This seems to happen several times a second and most likely results in a fairly even and accurate stir bar speed. It also results in the noise as the motor effectively accelerates and decelerates all the time.

The usage pattern is to put a vessel with liquid and stir bar on top, start the device and then adjust the speed such that you create a desired vortex. The exact RPM value is not really all that important. On the other hand, the temperature settings do need to be absolute and accurate. I think most users will vary the speed and eyeball the effect, then settle on what looks and sounds good. The LED indicators are a good indicator for eyeballing the relative speed and possibly being able to repeat a favourite setting.

I'm wondering whether a "close enough is good enough" approach to speed control could possibly result in a quiet mode of operation. I don't have enough information on the PWM implementation or choices, so can't make any concrete suggestions, but at a guess it should be possible to experiment a bit to see if there's some comfortable compromise between speed control and the noise generated by the motor constantly varying speed. I guess the ATtiny88 is probably not the right kind of device to attempt any sophisticated low pass signal filtering, but it should be possible to do something simple.

The LEDs do flicker a bit. It's actually when the external probe takes its temperature measurement.

I thought that would be the case. Given that it's not completely regular, my guess would be that the probe takes a variable amount of time to return a reading.

In the future we could try and talk to the probe by hacking the USART port but we'll need a new PCB design for that. Alternatively we could use a thermistor or analog temperature probe, but I really like that the DS18B20 is a known quantity and has great precision and accuracy.

Would it be possible to avoid hardware changes and use an interrupt handler to refresh the display? If the entire display refresh takes too long for an interrupt handler, would it be possible to handle only one LED refresh in each interrupt and have more frequent interrupts? Or is the problem caused by the probe and front panel sharing the same bus, thus disabling access to one device while waiting for another to respond?

Regarding the calibration, the temperature probe's spec is +/- 0.5˚C so the more you can insulate it from ambient the better. I think this is one of the main points where we can improve our design, either using a non-contact temperature sensor or supplying some kind of jacket in the future.

I'm pretty sure that the problem is not due to insulation or poor probe contact. Attaching the probe can get a little fiddly and the result does not have great aesthetics, but it is functional and flexible enough to suit a variety of vessels. I have now tried both probes on a Maverick ET-732 as well as an instant read thermometer and a mercury lab thermometer all located alongside the Yeast Forge probe. They were all placed right next to each other, under a large amount of bubble wrap and held in place by a rubber band. The Yeast Forge heater has held a reasonably consistent temperature, but always offset by +2C. I tried at 18C and 20C. Perhaps it's just an off-by-one error between the display and the actual setting, given that the display uses 2C increments.

 

[megabyte]

I think your description of the source of the sound is matching up with what I am hearing. So with the origin of the noise explained, I get to my usual bit where I say, "Geez, I wish I could get access to the firmware source code and start experimenting with the algorithms." I peeked inside and it looks like the six pin probe connector doubles up as an SPI in-circuit programming connector. I'd probably need to get a suitable programming cable, which at a guess would be a 5V interface and some kind of ATtiny88 development environment.

I love the open source/open hardware ethos, but in the years we've been manufacturing stir plates I haven't heard of much interest in hacking them. To get up and running you'll need an AVR ISP programmer or debugger and Atmel Studio but as you can see from the microcontroller choice, they weren't really designed to be hacked. If there's a wider audience that would like to be able to hack their stir plates we could use the ATMEGA328p in the next revision and port the code to arduino?

I'm wondering whether a "close enough is good enough" approach to speed control could possibly result in a quiet mode of operation.

Yes, I suppose there's a tradeoff between accuracy of speed and how abruptly the motor is controlled. The details of the PWM algorithm are that the motor output is recalculated every time a fan blade passes by the sensor (7 times per revolution). This is slightly different to most PWM algorithms that determine speed by counting the number of signal pulses in a fixed time period (which would average the differences in the blade widths). This was necessary to get the fastest possible loop response time so we can accurately control the motor speed down to 200RPM in the Yeast Forges (and even 100RPM in the digital stirplates) while handling varying load conditions. We're always looking for ways to improve so I'll keep your suggestions in mind.

Would it be possible to avoid hardware changes and use an interrupt handler to refresh the display? If the entire display refresh takes too long for an interrupt handler, would it be possible to handle only one LED refresh in each interrupt and have more frequent interrupts?

The display is multiplexed and uses interrupts to refresh it at a high frequency that tricks your eyes into thinking the LEDs are on at once, when actually it's only ever illuminating up to 4 LEDs at a time. The trouble is that reading from the DS18B20 is time critical so I have to disable the interrupts temporarily while temperature readings are being made which you see as a flicker on the display. Using the hardware USART register would mean the display updates could run unaffected.

I have now tried both probes on a Maverick ET-732 as well as an instant read thermometer and a mercury lab thermometer all located alongside the Yeast Forge probe. They were all placed right next to each other, under a large amount of bubble wrap and held in place by a rubber band. The Yeast Forge heater has held a reasonably consistent temperature, but always offset by +2C. I tried at 18C and 20C. Perhaps it's just an off-by-one error between the display and the actual setting, given that the display uses 2C increments.

I logged the temperature in the past (here) and it tracked around the target temperature very well, but it may be prudent to run another test in lieu of your recent observations. I'll collect a new dataset today.

 

[peteru]

I've been looking at the temperature readings a bit more.

I set the target temp to 18C last afternoon and checked the actual readings many times over about 8 hours. It stayed a fairly consistent 19.6C and in that time period I saw the heating LED illuminate on some occasions. When it did, it would turn on and off intermittently, for example on for 2 seconds, off for 3 seconds, on for less than one second, off for less than one second, on for 1 second, etc. The ambient temperature varied from 15C down to 12C. I checked again this morning and the temperature at that stage was 18.9C, but half an hour later it climbed to 19.1C.

I set the temperature to 20C at around 10:30 and by 12:00 it was sitting at 22.2C

So, the observations suggest that there is a consistent positive offset that ranges somewhere in the 1C to 2C range. I suppose it could be 2C if you give or take 0.5C for accuracy and another 0.5C for hysteresis.

 

[megabyte]

I've run some tests with my Yeast forge set to 20˚C, I set it up shortly after my last post as promised. It reached its target temperature at 12:34pm and has overshot the marker by 0.25˚C which is within spec of the temperature probe's +/-0.5˚C. The graph seems to have stabilised as expected. It's still collecting data and you can view it live here. Forgive the peak in my room temp, I had turned the heater on because I'm freezing here but I turned it off again before it got high enough to affect the experiment

Here are pics of my test setup. External thermometer showing 20C and the internal one is showing just a smidge over.
Test setup.


External temp.


Internal temp (a bit hard to see with condensation).


I'm not sure where to go from here since everything still tests out ok at my end. Naturally if you think there's an issue with your yeast forge I'll be happy to check it out under warranty. Do you have any ideas?

Edit: Do you have a starter in there that could be heating it?
Edit: It's been sitting stable at a final value of 20.31˚C for 30 minutes so I'm going to stop collecting data.

 

[A.B.]

As a data point, I got my unit from Clever Brewing. My fan isn't noisy, but I have noticed if I hold the "down" on past 2 beeps I have to do a power cycle. No problem really to me.

 

[peteru]

I've run some tests with my Yeast forge set to 20˚C
...
I'm not sure where to go from here since everything still tests out ok at my end. Naturally if you think there's an issue with your yeast forge I'll be happy to check it out under warranty. Do you have any ideas?

Edit: Do you have a starter in there that could be heating it?
Edit: It's been sitting stable at a final value of 20.31˚C for 30 minutes so I'm going to stop collecting data.

Clearly your configuration appears to do a very good job in the experiment. My Yeast Forge is also doing a good job of maintaining a steady temperature, it's just that the absolute value has a 2C offset. Luckily it's not too hard to mentally compensate for that - I just need to set my Yeast Forge one step below the temperature I actually want. It's not a show stopper for me, but if you are like me, you'll probably want to get to the bottom of this.

It could be down to a bad / marginal probe - unlikely, but not impossible. If you have a spare probe and want to pop it in an envelope and post it, I'll swap the probes over and see if it makes any difference. It's probably the simplest first step.

I have a 1L starter in there now. Before that, it was a 500mL yeast step up. I'm limited by the size of my vessels. It also occurred to me that the yeast activity may have been driving the temperature high, so I turned off heating and continued to stir at the same speed. The temperature dropped from 22.2C to 20.6 in the space of less than an hour. I set the Yeast Forge back to 20C target and it climbed to 21.8C in about 1/2 hour. I'd say that this pretty much proves that the overshoot is due to the YF heater.

 

[megabyte]

Yep I think we're on the same page, it's either our probes or our test setups that differ.

My test setup worked great, but that's expected I guess because this is pretty much how it was designed and tested. I've done everything possible to make sure I'm measuring the same spot that the Yeast Forge's probe is measuring. I'm using a large 2L flask with a nice surface to lay the probe flat against and I've got a decent amount of overlap with the bubble wrap. I've also mounted the yeast forge's probe parallel to, and in contact with, both the digital logging probe and my external thermometer to ensure as much as possible that they're not measuring different things.

As you can tell, it's bugging me because I have a lot of faith in the DS18B20 and the Yeast Forge. If I had your yeast forge with my test setup, would I get the same test results? I don't know but I'd like to find out. I'll PM you and see if I can pry it away from you some time or post you out that spare probe