TB6600 drive from EBAY

Just got done reversing the latest batch of Ebay TB6600 cards.
This one is marked BL-TB6600-V1.2 D412
Note in the photos the one without the heatsink I moved all the tall components to the back side so I can mount it to a larger heatsink assembly.
[ATTACH=CONFIG]330182[/ATTACH]
[ATTACH=CONFIG]330184[/ATTACH]
[ATTACH=CONFIG]330186[/ATTACH]

And the reversed schematic
[ATTACH=CONFIG]330188[/ATTACH]

The sort of good:
They used a 74HC123 retriggerable monostable multivibrator with a time constant of ~0.45s.
- When a pulse comes in (led on) it triggers the timer and drives the fet off and stays off till timer expires.
- When the pulse is done (led off) the driver steps and as long as the pulse is shorter than the timer it is still in full current mode.
- Every off to on transition of the led causes the timer to extend the timeout. Other than dropping power there is no way to get a runt timeout.
- This does not drive the TQ pin but rather puts a 10k resistor in parallel with the variable resistor. With the current setup this results in ~45% hold current at full range and ~55% at mid range. (may be an issue with small steppers with low current limits. But if that is the case pick a smaller driver.)
- Also this function can be disabled by turning SW4 off. Then the driver will run in full current mode all the time.

The Unknown:
Rosc is a 27k resistor. Data sheet recommends a 51k. Not sure what this would affect or if 27k is too low. (can anyone comment on this?)
Enable is active (driver running) with no input. (could fix by rewiring the opto output but don't care about in my application as EPO takes out the motor supply.)

The Bad:
1.) no thermal grease (seems like the norm. Also had one board with a missing screw that was broken off and another with the 470uF cap with one leg ripped out)
2.) Vref can exceed 1.95V (to fix replace R11 with a 31.6k resistor)
3.) AL output overloaded (to fix remove the AL led and replace with a 30.1k resistor)
4.) With Vref fixed max output is only 3.1A (to fix add a 0.47 ohm 1% 1W resistor in parallel with both banks of sense resistors will allow full 4.5A output)
5.) MO output overloaded (to fix remove the RUN led and replace with a 30.1k resistor)
6.) 5V reg overloaded (to partially fix remove the power led and add a 0.1uF cap from pin 24 to ground near pin 24)

I think that is everything on this board. Have one more board of a different design I need to trace out. Hopefully will have it's results yet this week.

Second card in this batch.
Marked MIB JY6600SW
Note I removed the heatsink before taking photos.
[ATTACH=CONFIG]330290[/ATTACH]
[ATTACH=CONFIG]330292[/ATTACH]
[ATTACH=CONFIG]330294[/ATTACH]

The sort of good:
Dual 6N137s on both step and dir.
Vref circuit correctly limits voltage to 0.3V to 1.95V ish.
Rosc in the valid rang (51k)

The unknown:
Enable is active (driver running) with no input. (could fix by rewiring the opto output but don't care about in my application as EPO takes out the motor supply.)
Power reduction circuit drives the TQ pin so output will always be either 100% or 30% with no disable switch. Could be modified to disable (tie TQ pin high) or change the ref point (for 50% reduced power use a 7.32k and another 2n7002 Nfet. see previous card design for rough hookup design)

The Bad:
1.) Heat sink glued on with an elastomer (may be heat sink compound may just be kitchen and bath) but was done tipped so large filled gap along one side.
2.) No filter cap on Vref (to fix add 0.1uF to line)
3.) AL output overloaded (to fix remove the AL led and replace with a 30.1k resistor)
4.) With Vref fixed max output is only 4.3A (to fix replace the 0.56 ohm resistors with 0.47 ohm 1% 1W resistors this will allow full 4.5A output)
5.) MO output floating (to fix add a 30.1k resistor from pin to +5V)
6.) 5V reg overloaded (to partially fix remove the power led, work led, lift U7 pin 3 and ground it, and add a 0.1uF cap from pin 24 to ground near pin 24)
7.) No filter cap on clock line (to fix add 100pF cap to line)
8.) Floating leftover gates in the 74VHCT14 (to fix tie the inputs of the unused gates to ground)
9.) Reset pin tied to power. (to fix lift pin and add a 12k resistor to +5V and a 0.1uF cap to ground)
10.) Missing filter caps on input power rail (to fix add two 0.1uF caps near the power pins of the TB6600)
11.) Missing input power protection diode (in theory not needed but be extra careful not to hook up backwards)
12.) Time constant of the reduce power circuit very short at 0.05s (to fix replace C1 with a 6.8uF cap and replace R19 with a 100k resistor to get a 0.68s TC)

wireb,

I saw your post above about the TB6600 drives from e-bay, and was wondering if you could help me out in regard to setting driver current on these drives. It appears that the 2 pads indicated by the arrows in the picture below are for setting the Vref which determines the output current of the drive. What I don't know is what Vref voltages at those pads relates to what output current. Can you help with a formula or something?

[ATTACH=CONFIG]340540[/ATTACH]

Yes there is a formula but need some more info off your board. On the other side there should be some large sense resistors. Need to know the values of these resistors (number code on them) and how they are arranged to fill out the formula.

First the formula.
Io (100%) = (1/3 × Vref) ÷ RNF
where Vref is the voltage at the variable resistor
and RNF is the value of the sense resistor combination.

Also note there are some limits you need to be within for proper operation (many of these cards a miss designed and exceed these which can cause issues.)
0.11Ω ≤ RNF ≤ 0.5Ω,
0.3V ≤ Vref ≤ 1.95V

Post a photo of the resistors and I can get you the RNF value you need for the formula above.

Thanks. I will get you a picture as soon as I can.

Here is the best picture that I could get with my phone. Hopefully it has what you need. If not, let me know and i will get a better camera.

So I think I have the Vref figured out for this drive I have. My board has three 0.62 ohm resistors in parallel for the sense resistors. So the Rnf value would be 0.206 ohms. Then using the formula for the output current:

I=(0.333*Vref) / Rnf

with a little re-arranging:

Vref = 3 * I * Rnf

So for my application with 2.8 amp per phase steppers, I need Vref to be 1.736 volts. Since this doesn't exceed the 1.95v limit I'm OK there, and I can get the current I need.

For those that are like me and don't know much about electronics, the sense resistors are shown in the attached picture. In my case there were three 0.62 ohm denoted by the R620, and the formula for parallel resistance can be found on the internet, or there are calculators too.

[ATTACH=CONFIG]344224[/ATTACH]

Yep that is within a tenth of what I calculated and your correct that Vref is in the "safe zone" for the input. Another side note for the sense resistors there are 2 sets on every card. The 3 on the top are for the A channel and the 3 on the bottom are for the B channel. For the calculations just use one set (not both). Both sides should be the same value else very weird things will happen.

Sorry I totally missed your previous post else I would have responded sooner.

I found [URL="https://www.amazon.com/gp/product/B016ZJS1FA/ref=ox_sc_act_title_1?ie=UTF8&psc=1&smid=A2OCOGC9B25845#Ask"]this[/URL] "UPGRADED VERSION" on Amazon with 5 star rating from all 13 purchasers. Then I looked at ebay and found the upgraded version for under $8. I searched this thread and the forum for the "UPGRADED VERSION" and found no reference. Trying to make sense of it. I was ready to skip the 6600 and get the 542, but there seems to be 100% happy users with the upgraded version i found. What say ye electronics masters?

[QUOTE=mikeysp;1980290]I found [URL="https://www.amazon.com/gp/product/B016ZJS1FA/ref=ox_sc_act_title_1?ie=UTF8&psc=1&smid=A2OCOGC9B25845#Ask"]this[/URL] "UPGRADED VERSION" on Amazon with 5 star rating from all 13 purchasers. Then I looked at ebay and found the upgraded version for under $8. I searched this thread and the forum for the "UPGRADED VERSION" and found no reference. Trying to make sense of it. I was ready to skip the 6600 and get the 542, but there seems to be 100% happy users with the upgraded version i found. What say ye electronics masters?[/QUOTE]

Unfortunately, unless someone has taken the time to analyze the particular TB6600 driver you are referring, there is no way for anyone to know how good or bad it is. They all work to some degree, but how well depends on how well designed the circuit is. This pretty much requires that someone that knows electronics either analyze it on an oscilloscope, or examine the circuit itself.

I can say that using the TB6600 drivers I have, which are like the one in the attached picture, that they work fine and I have had no issues with them. However, like the other TB6600 drivers, they have some things that aren't quite right, but for my purposes and the steppers I am using they are fine because the problems are not relevant for the motors and motor current I am using. If I wanted to get full power from the driver, I would have to make some modifications, but since I am only needing 2.8 amps, the modifications aren't mandatory.

[ATTACH=CONFIG]344312[/ATTACH]

JB109, thank you for your response. I spent the night reading [URL="http://techref.massmind.org/techref/io/steppers.htm#Estimating"]MassMind's[/URL] info on stepper/driver selection. From what I learned there, combined with your response, I am deducing:

1. The TB6600s found on ebay often work fine if your amperage requirements are not significant (2.8amps in your case); however, if you want the 4+ Amps that the chip is capable of producing, you have to find someone who makes the driver properly, or make the modifications to the poorly designed driver.
2. In my case, I am driving a 40lb +/- gantry for my 5x10 plasma table; and, according to [URL="http://techref.massmind.org/techref/io/stepper/estimate.htm"]this calculator[/URL] if I want rapids of 600ipm, I only need 45 Watts.
3. Since I am using a belt drive (direct) with a 15T, 5.08mm pitch pully, which means 5.08 x 15 = 76.2mm (3.0") of movement per revolution of the stepper, so I will need to operate at 200rpm (3" x 200rpm = 600 IPM) and the steppers I am looking at are ebay Nema 23, 270oz in (190 cm-N), 3 Amp, 3.5mH Inductance, 4-Wire:

So, according to the Estimating Maximum Speed calculator on the same page, if I operate Bipolar-Serial I will achieve 84 Watts (nearly double my need) at 199.8 RPM; however, the minimum time per step will be 1.5ms. I don't understand importance of time per step yet, so I do not know if this is a problem. Anyways, it appears that for an application like mine, I would do well with the TB6600 without any mods, predicated on the driver working when it arrives.

I feel like a guy with two days of kung fu training. Enough to get beat up. However, I would rather take my licks on the forum than in my pocket book. :)

Mikeysp,

The particular TB6600 that I have would not exceed any of the parameters on the TB6600 datasheet up to about 3.1 amps per phase. In my application at 2.8 amps per phase, they seem to work pretty well, but I can't really compare it to anything. I have another machine with better drivers and motors, but that machine has ball screws, whereas the one with the TB660 is acme screws. So I can;t really compare performance because it is apples to oranges.

Keep in mind that it really does depend on which TB6600 you buy. All of the different drives seem to have different problems. I kind of lucked out because this particular drive seems to have fewer problems and the problems it does have could be pretty easily fixed. Not so much on some of the other ones.

I bought these because they were $6 each and for $18 for three I figured it was worth the gamble. It paid off because I get enough performance on my machine to satisfy me, but I wasn't going for super high performance. This is a smallish dovetail milling machine table with 0.1" pitch acme screws, belt reduction drives, and cheap steppers at 24 volts. I can get 40 IPM rapids, which is about what I was hoping for. At that speed the steppers are running 600 rpm. I could probably get more performance with a 40 volt supply and better motors, but this was a low cost exercise.

As for the 1.5milliseconds step time number, I can explain that for you. You are looking to achieve 200 RPM, and the motor is 200 steps/rev, so 200 x 200 = 40000 steps/minute required, or 40000 / 60 = 666.6 steps/second Then 1/666.6 = 1.5 milliseconds.

However, this doesn't include any microstepping, and since you are using a belt drive, you will want to use microstepping. I personally would use 1/16 microstepping for a coarse belt drive like that, which means that you controller would need to supply 200 * 16 = 3200 step pulses to move the motor one revolution. So now if you re-run the calcs with that in mind then you get a step time of 94 MICRO-seconds, and a pulse rate of 10.6kHz. Those numbers aren't particularly fast and any controller should be able to handle that as well as the TB6600 drivers. The TB6600 can take 200kHz pulse train.

Since you only need 200 rpm then these drivers may be a good choice. If you are willing to risk $18, I would say give them a try. I would use a 36-40V power supply with them because the absolute max for the chip is 50 and 36-40 would leave some headroom. I am happy with mine.

I have that same driver that 109jb has pictured. Well atleast it looks exactly the same. Bought in eBay for similar price. I tested it unmodified with about a dozen different stepper motors ranging from nema17 to nema34 size (50oz-in to 450oz-in). Low inductance to high inductance motors. 24-36volts, current adjusted per motor but 3.5amps setting was max I used. My unscientific tests concluded that the lower inductance motors ran very well. Hardly any resonance issues and pretty high rpm. Higher inductance ones had noticeable resonance problems. The motors that were 2.5mh or less ran the best. Just my observation if you end up using a tb6600, get lower inductance motors.

Edit I should also add that it didn't really make much difference changing power supply voltage or the size/torque of the motor as far as resonance. The ones with the low inductance seem to have less mid band resonance issues. I have a really really low inductance nema34 superior electric stepper. The Tb6600 spun that motor well over 1000rpm with no noticeable resonance in the 3-400rpm region.

109JB, thank you for such an instructional response. Most helpful. I am almost certain I am going to take the risk and try the TB6600 on the plasma build.

Jfong, thank you for sharing your test results. I had considered these 2.4mH steppers. They are about $89 for 3 delivered. I had hoped the 3.5mH versions (3 for $65) would have been good to go; however, with all your testing, you make a good case for lower inductance versions. I will do some hunting on ebay, etc.. to see if I can find some lower inductance versions..

You have to love the information sharing the internet has provided and generous people like yourselves have contributed to.

-Mike

jfong, I have done a ton of reading on the effect of inductance since you shared with me the tests you ran. Question: while you stated that below 2.5mH your stock TB6600 did great, I have discovered that there is a huge spectrum 2-18mH or more. Did you test a 3.5mH stepper? If yes, how did it do?

Thanks,

-Mike

[QUOTE=mikeysp;1982196]jfong, I have done a ton of reading on the effect of inductance since you shared with me the tests you ran. Question: while you stated that below 2.5mH your stock TB6600 did great, I have discovered that there is a huge spectrum 2-18mH or more. Did you test a 3.5mH stepper? If yes, how did it do?

Thanks,

-Mike[/QUOTE]

I did this test awhile ago. I would have to look at my notes to see. With any driver, a lower inductance gives potentially higher rpm. I mostly use anti-resonance drives from companies like Parker and Geckodrives so I am not overly concerned about it. The TB6600 is more susceptible to resonance and I have noticed that running lower inductance motors are less likely to have that issue. Unless I have the exact same stepper motor you plan to buy, it may exhibit different operating characteristics with other 3.5mh motors. I have many different stepper drives/motors collected over a span of 20 years. They all seem to behave differently. Some drives work much better than others.

I purchased [URL="http://www.ebay.com/itm/222353534440"]these[/URL] "UPGRADED" TB6600 32 segments 4A 9-42VDC and some 3.5mH 270in oz Nema 23 Steppers.

On a good note, the heatsink did have the paste. However, being unsure what upgraded means, can anyone tell anything good or bad from the pics? Once I figure out how to connect them up and how to test performance, I will plan on posting the results.

[ATTACH=CONFIG]347066[/ATTACH]
[ATTACH=CONFIG]347068[/ATTACH]
[ATTACH=CONFIG]347070[/ATTACH]
[ATTACH=CONFIG]347072[/ATTACH]

Thanks,

-Mike

Not sure that is a TB6600 driver have not seen a TB6600 in a BGA package before. Can you carefully clean of the heatsink compound from the top of the IC and post the part numbers marked on top?

wireb, the info from the IC is:

S109AFTG
164IH
D71774
JAPAN

Here are a couple of pictures:

[ATTACH=CONFIG]347116[/ATTACH]
[ATTACH=CONFIG]347118[/ATTACH]

Not TB6600

[url]https://toshiba.semicon-storage.com/info/docget.jsp?did=14642&prodName=TB67S109AFTG[/url]

Need to order one to see how well it performs.