Auto Bed Leveling troubles

[Roxy]

I found out Makerfarm uses a different thermistortables.h file than the newer versions of Marlin. Copying my old thermistortables.h file from my original Markerfarm configured version of Marlin over to the new version of Marlin with bed auto leveling fixed my problem with the bed not heating up right.

With original RAMPS firmware I had no problem getting above 90. With Marlin I have the bed set at 100 for ABS but it stops at like 87

I think it would be wise to check the ThermistorTables.h file and see if you have similar values now compared to what used to be there. Do you know the constants for your thermistor? I suspect if you change your thermistor type the problem can be remedied.

[Pronus]

From what I understand Makerfarm I3 and I3V's use a 100k EPCOS thermistor. It even says in the configuration.h file (100k EPCOS - Not as accurate as table 1 (created using a fluke thermocouple) (4.7k pullup)

I remember figuring out the problem when I had my original firmware from Markerfarm loaded and I had my bed turned up to around 90C. I turned off the printer and quickly uploaded to the newer version of Marlin with the bed auto leveling. When I turned my printer back on the bed was actually reading at least 10 degrees higher than before I even turned off the printer. I knew there was no way I turned off the printer for about a minute and the bed actually got hotter lol.

As soon as I copied over the thermistortable.h file to my newer version of Marlin, it totally fixed my issues.

[Roxy]

As soon as I copied over the thermistortable.h file to my newer version of Marlin, it totally fixed my issues.

The only bad thing is you don't want to be copying old files into new releases all the time. I bet there is a thermistor type in the new file that accurately models what you have. If that could be figured out, it would be safer and cleaner to just change the thermistor types in Configuration.h to the type that is correct. It kind of sounds like the type declared for the Makefarm isn't right ???

[dacb]

IIRC, table six is different. I don't copy the table over, I have a diff that I patch in to fix the table, but that's the best I've been able to find. Here is my diff with that table: http://3dprintboard.com/showthread.p...ll=1#post26056

[Roxy]

You are using a Type-6 Thermistor??? Can you post the old? I went through this a little bit ago because my thermistor broke one time I was unclogging my nozzle. I ordered some thermistors off of eBay but could not get exactly what I needed at first. So... I ended up having to learn about all the constants and how the table is constructed. If you post the old (working) and confirm it is a Type-6 Thermistor I'll take a look.

I just compared the Type-1 to Type-6 which are supposedly for the same thermistor. There is a lot of deviation once you get above 150 C.

[dollarz81]

Everything is working great. I really appreciate all the help.

[TopJimmyCooks]

Going from memory here, but Zennmasterm in one of his videos about upgrading marlin for a makerfarm i3 mentioned that thermistor table 6 is what comes with the stock makerfarm firmware based on the supplied currently shipping thermistors, but when upgrading to newer Marlin it is safe to switch to table 1. Check the video though.

[printbus]

The Type-1 vs Type-6 issue here is, IMO, a classic example of what can go wrong in open source firmware where people can make whatever changes they want.

1) Both type 1 and type 6 are for the same thermistor type. There's a comment in the configuration.h file that reads "// 6 is 100k EPCOS - Not as accurate as table 1 (created using a fluke thermocouple) (4.7k pullup)". Why are there two definitions for the same thermistor? Someone wants to use the less accurate one?

2) Comments in the code are unclear. Does the "created using a fluke thermocouple" apply to the type 6 data or is it explaining why the type 1 table is better? If it applies to type 6, it's hard to believe it would be less accurate than type 1. If it is explaining why type 1 is better, why in the heck isn't the comment placed with the type 1 data?

3) Somewhere along the way, MakerFarm picked up data for Type 6 that differs from the Marlin baseline. Dacb has posted the diff info. Is it just old data and the "not so accurate" type 6 data has been improved in newer Marlin versions? Did MakerFarm (or their supplier) tailor the table data for Type 6?

4) Assuming MakerFarm had some reason to modify the data for Type 6, comments in the code explaining the change would have been nice. Same as it would have been nice for other places MakerFarm tailored things to have been commented. Without comments, we have no way to know whether the changes are because they're using something other than an EPCOS, had some test data to indicate they could improve on the "not so accurate" Type 6, or had some reason to want temperatures to be "scaled" from reality.

<rant off>

[TopJimmyCooks]

I agree with this rant. ^^^

The only thing that I might add is that the epcos 100K thermistor seems to be very widely used and well understood. Should be some consensus on how to table it.

Another item: makerfarm sells their own version of the prusa version X bed heater pcb, perhaps changes were made to better reflect measured temperatures of this heater. would be nice if that data was made known.

[Roxy]

1) Both type 1 and type 6 are for the same thermistor type. There's a comment in the configuration.h file that reads "// 6 is 100k EPCOS - Not as accurate as table 1 (created using a fluke thermocouple) (4.7k pullup)". Why are there two definitions for the same thermistor? Someone wants to use the less accurate one?

It was a while ago... I don't remember everything I went through (and learned). But I think the answer is close to this: Depending upon what pull up resistor value you use, you get a table with different numbers. While Type-1 and Type-6 are for the same thermistor, my guess is they were generated with different pull up resistor values and as a result, you get different numbers in the table.

I don't know that this is true, but if so... It should be much better documented so this kind of confusion doesn't happen.