[HJozsi]

I posted jozsi's video on youtube, I hope its ok with u Jozsi.
.
.
for all!

RGDS
Irfanulla

OK, no problem...

I'm sure I'll get some discout from the next kit ....



Regards,
Jozsi

[contactirfu]

OK, no problem...

I'm sure I'll get some discout from the next kit ....



Regards,
Jozsi

You bet! I have to take notes now on the changes I have to do myself. once i get bk to my machine!

RGDS
Irfan

[kreutz]

Thank you Kreutz!
I will replace C28 with a tantalum. Since the voltage across the sense resistor is in direct relation to motorcurrent shouldnt the p-p voltage across the sense resistor be lower when the slow limit activates?

I'llplay with this a bit more and see what I can come with it.

/Henrik.

The peak voltage across the sense resistor will be a little above the set voltage on the Test point (once the LED lights ON), how much above depends on the chopper circuit delay and motor electrical time constant for a given power supply voltage.

[rbond]

Hi All
Just wondering how the testing is going on the new hp uhu drives?

Rick

[H.O]

Hi,

I haven't replaced C28 on my board yet but I've verified that it operates at the peak-limit all the time. While applying external torque to the motor and adjusting the trim-pot the peak voltage across the sense resistor tracks Vref_OC perfectly, no matter how long you apply torque to the motorshaft.

Surely Jozsi and me can't be the only one trying them out - or is that the case? If someone else has a drive, can you please check how the slow limit performs on your board. Does the yellow LED turn ON? If so do you see/hear/feel any difference on the motor?

Joszi, have you found the reason why you had to set your Vref so high yet?

/Henrik.

[HJozsi]

Hi,

I haven't replaced C28 on my board yet but I've verified that it operates at the peak-limit all the time. While applying external torque to the motor and adjusting the trim-pot the peak voltage across the sense resistor tracks Vref_OC perfectly, no matter how long you apply torque to the motorshaft.

Surely Jozsi and me can't be the only one trying them out - or is that the case? If someone else has a drive, can you please check how the slow limit performs on your board. Does the yellow LED turn ON? If so do you see/hear/feel any difference on the motor?

Joszi, have you found the reason why you had to set your Vref so high yet?

/Henrik.

Hi Henrik,

I made some additional changes according to Kreutz advise because the fast limit is two time higher than the slow and I set slow to 15A... The additional change was C28 take out, and R20 take out. When R20 was out I had to set the Vref higher in order to have tha same power than before ... I have put it back, so I set 0.15V to 15A current limit. C28 I replaced with 3.3uF tantalium but it seemd no chance to hit the Ref1 level ... I turned also down to 0.03V the Vref ( current limit ) and turned the the motor axis and was hold for half a minute it was far from ... Then I soldered a 10kohm parallel to R21, still was not enough, then one more 10kohm. In that case the led turned on according to C28 R40 time constatnt, and the axe become weaker a bit and changed it's voice to high tone. I did not spend too much time with that, because when I turned the current higher to about to 6A and tried to test the slow current limit, something happened when slow current limit turned on. The motor become weak and it remained for those direction. The opposite direction was ok ( strong) . I Thought, a fetdriver, or Fet gone away. Turned off, turned on, the gates had good signal. The motor was standing but still one direction was strong, the opposite was weaker. I put the scope head to U9 pin 1 and the motor become strong again... releasing weak ... Then I replaced U9 with a new one and that solved the problem... Finally I left only one 10k paralel to R21. Interesting ... It seems to me the fast limit can be active only and determines the power of the motor... Tomorrow I'm going to test with the lathe again, and if it works well I have to leave there ...
So I can analyse the current limit, when I order my next kit for the next lathe ..

Regards,

Jozsi

[kreutz]

Hi,

I haven't replaced C28 on my board yet but I've verified that it operates at the peak-limit all the time. While applying external torque to the motor and adjusting the trim-pot the peak voltage across the sense resistor tracks Vref_OC perfectly, no matter how long you apply torque to the motorshaft.

Surely Jozsi and me can't be the only one trying them out - or is that the case? If someone else has a drive, can you please check how the slow limit performs on your board. Does the yellow LED turn ON? If so do you see/hear/feel any difference on the motor?

Joszi, have you found the reason why you had to set your Vref so high yet?

/Henrik.

Hello,

I am going to re-check the behavior of the slow limit circuit, on my first prototype, as soon as I go back home next week. Thank you for your help on the tests!!

Regards,

Kreutz.

[H.O]

Jozsi,
That is VERY interesting because now mine behaves exactly like yours - strong in one direction and sluggish in the other. It's not completely dead like it was when my PCB was shorted but it just feels soft/slugish in one direction, exactly like you describe. I'll have a look at U9 when I get a chance.

Kreutz,
No worries, we all know you're on the road - I was hoping someone else also had a drive so we could confirm (or not) what is going on.

Thanks!
/Henrik.

[contactirfu]

U9 is LM339N right? what happening with that?

[kreutz]

Hello,

I am going to re-check the behavior of the slow limit circuit, on my first prototype, as soon as I go back home next week. Thank you for your help on the tests!!

Regards,

Kreutz.

Please, try a new test replacing R21 by 4.70K 1% 1/4W resistor and C28 by 4.7 uf /25v tantalum. The new reference voltage will be 2.97V.

Kreutz.

[vrsculptor]

Advanced Motion Controls (a-m-c.com) has very reasonable prices on shunt regulators for servo systems. Their srst175 (175 volt, 95 watt) regulator is ~$100 quantity one. They have them in other voltage ranges.

The price seems very reasonable for a commercial CE rated part but I don't know how to judge if 95 watts provides enough protection. Henrick's design could handle much more power. If 95 watts was enough I think I would rather buy an assembled unit as I already have enough irons on the fire.

SPECIFICATIONS FOR SHUNT VOLTAGES UNDER 200 V
FUSE 3 A Motor Delay rated @ 250 VAC
FILTER CAPACITANCE 1200 mF
DISIPATION CAPABILITIES 95 W
RESISTANCE 5 W
SIZE
8.00 x 4.25 x 2.63** inches
203.2 x 108.0 x 66.7** mm
WEIGHT .32 lbs.
.14 Kg.

PS. Received Uli's chips Friday. Hope to assemble UHU next weekend.

[eidos]

hi boys

this is my version of the UHU servo .
I call it hybrid UHU driver.
in fact I have founded on the original driver of Uli .
but the power section has been modified in accord to how much realized by Kreutz end other boys.
I have also modified the out stage section (snubber) of the motor.
removed the two spools / inductances, in as for my notice they can have put out of the pcb when needs.

I have realized the section of RS232 interface on a separate card.
what I can connect when it serves.

I would like to ask you a thing.
according to you, there are negative effects to leave disconnected by the Micro the two lines (pin 2 + pin 3 ) that go from the section of interface RS232 ? .

hi
good milling to everybody

[kreutz]

hi boys

this is my version of the UHU servo .
I call it hybrid UHU driver.
in fact I have founded on the original driver of Uli .
but the power section has been modified in accord to how much realized by Kreutz end other boys.
I have also modified the out stage section (snubber) of the motor.
removed the two spools / inductances, in as for my notice they can have put out of the pcb when needs.

I have realized the section of RS232 interface on a separate card.
what I can connect when it serves.

I would like to ask you a thing.
according to you, there are negative effects to leave disconnected by the Micro the two lines (pin 2 + pin 3 ) that go from the section of interface RS232 ? .

hi
good milling to everybody

Theoretically there should be no problem by leaving the Rx and Tx pins unconnected. On the real world, nevertheless, noise captured on the Rx pin could trigger interrupts and overflow the software stack on the UHU chip leading to unexpected behavior or resets.

On the suppression of the inductors: They are there for a reason, the reason is to limit the discharge current of the snubber capacitor ( di/dt) when the Mosfet turns on, so protecting the Mosfets, and also limiting switching losses. The RLC filter's function is primarily to protect the motor against high dv/dt transients. As you said, it might work without it, but I can't guarantee it will work for a long time. I assure you I am not a shareholder of the inductors manufacturer's company .

Thanks for sharing your PCB design.

Kreutz.

[eidos]

Theoretically there should be no problem by leaving the Rx and Tx pins unconnected. On the real world, nevertheless, noise captured on the Rx pin could trigger interrupts and overflow the software stack on the UHU chip leading to unexpected behavior or resets.
cut....

I am not a shareholder of the inductors manufacturer's company .

Thanks for sharing your PCB design.

Kreutz.

hi kreutz
not even I am not a seller of inductances.
however these occupy a lot of space on the pcb.

my idea is this:
is it possible to get the same effect of the inductances using some diodes?
you see the datascheet of ST L6203.
does thing think of it?

hi
Salvatore

[kreutz]

hi kreutz
not even I am not a seller of inductances.
however these occupy a lot of space on the pcb.

my idea is this:
is it possible to get the same effect of the inductances using some diodes?
you see the datascheet of ST L6203.
does thing think of it?

hi
Salvatore

Salvatore;

There is a lot of difference between a low current / low voltage design and a high current / high voltage design, there are different rules for each kind of design.

Any diode handling 25 amp current is going to introduce high power losses and require a heat-sink. The power stage Mosfets already have intrinsic anti-parallel fast diodes.

The UHU HP version is intended to be low priced, I introduced the RLC filter in order to protect the investment of the DIY person assembling it. Yes, it requires a little more board space but that is compensated by the savings, if they are not used, the power handling capacity of the board is going to be limited, which would require oversizing the output stage Mosfets and even reducing switching frequency (not possible without tampering with the UHU chip's firmware) in order to overcome.

If you have to replace the servo motor because of insulation failure due to high dv/dt you will probably expend more money than the cost of the UHU HP controller and a lot more than the cost of the inductors. 185v @ 25 Amps handling is more than 4.5 KW. That power handling does not belong to a little controller board anymore.

Even on the safety related features there are a lot of differences, while less than 48 volts are considered relatively safe to handle (depending on current), a board like this one, handling high voltages and currents, is potentially lethal in case of an accident. High power motor loads (moving) have (potentially) a lot of mechanical energy and are also dangerous.

I think the UHU HP board is really squeezing to the limits the capabilities of the UHU chip, which was not designed with the high power handling safety in mind (and have surprising features for an 8 bits micro-controller with only 2K Flash and 128 bytes RAM). That is why I added an isolated RS232 interface on board, as well as a missing encoder cable safety, and replaced the power connectors. There are more safety related improvements I would like to see implemented, but they will keep on adding to the cost of the boards. I also thought about using a modular approach for the RS232 interface but decided otherwise because, in order to safely connect it, it would be necessary to power down the board.

Regards,

Kreutz.

[tenmetalman]

Hi,
Irfanulla & myself have decided to release the bare PCB's due to so many requests. the price is: one or two PCB's $25.00us each, Three PCB'S as a package deal $65.00us. This is plus shipping. If interested please go to our site:
[email protected]
Thanks
Paul (tenmetalman)

[eidos]

Salvatore;

There is a lot of difference between a low current / low voltage design and a high current / high voltage design, there are different rules for each kind of design.

Any diode handling 25 amp current is going to introduce high power losses and require a heat-sink. The power stage Mosfets already have intrinsic anti-parallel fast diodes.

cut.....

Kreutz.

Hi Kreutz
we go for order:
the function of the snubber is that to limit the transients produced by the motor winding when the driver commutes the polarity of the phase, correct ?
is it the motor that produces HV transient then, not the H bridge of the driver, correct?
the function of the diodes for as described on the datascheet St, it is that to make ricirculating the current between GND and the motor phases,
to limit the oscillations of tension on the GND plain.

in Your case (your RCL filtering circuit ) the concept to my notice is wrong.
in how much through the inductances she limits the current in exit toward the motor, obviously filtering the HV spikes with the snubbers toward " GND LOOP ".
but to my notice it is wrong to send the motor extratensions to GND, the snubber it has alone sense if connected among the " two poles of the motor" .
two inductances can be left, in effects they serve to limit the HV spikes but you must be put after the snubber and in exit toward the motor.
for as described in attached figure.
now I don't know what you have seen with your oscilloscope.
or as has effected your measurements.
but I believe that there is a reasonable doubt on thing both the cause of the troubles that you have seen, and on as them must be corrected.
my opinion.

best regards
Salvatore

[kreutz]

Hello Salvatore;

the function of the snubber is that to limit the transients produced by the motor winding when the driver commutes the polarity of the phase, correct ?

No in this case. You are comparing a slow low voltage H bridge output with a high voltage high speed one, in the L6203 case dv/dt is not an issue due to slow switching times and low power supply voltages, in our case, we don't slow down switching times on the Mosfets (our Mosfets are a lot faster), and dv/dt is too high for the motor, so the RLC snubber network is used in order to control dv/dt at the motor terminals.

the function of the diodes for as described on the datascheet St, it is that to make ricirculating the current between GND and the motor phases, to limit the oscillations of tension on the GND plain.

ST use extra diodes because the internal anti-parallel diodes' recovery time is too big (Using slow reverse recovery diodes does not help controlling high Drain-source voltage surges, when switching off inductive loads, during the reverse recovery period). We use the Mosfets' intrinsic diodes that are fast enough for the application, also our Mosfets have a clamping voltage capability with enough energy handling to limit those pulses if present.

In order to limit parasitic inductances we designed the PCB traces very carefully, we also use separated ground planes for the logic and power stages joined at only one point.

in Your case (your RCL filtering circuit ) the concept to my notice is wrong.

We all have the right to be wrong once in a while , meanwhile you can go to the "International Rectifier" website and take a look at the application notes, there is one for designing the same kind of filters for 3 phase motor drives (Design of the Inverter Output Filter for Motor Drives with IRAMS Power Modules), maybe they are also wrong... anyways, I am always open to learn...

I believe that there is a reasonable doubt on thing both the cause of the troubles that you have seen

So far the problems found on the Rev1.0b prototypes have been related to the simple slow/fast current limiting circuit being employed above the intended design currents (more than 11 Amps continuous stall current). I already posted a solution: Eliminating the Slow/Fast operation by replacing C28 by 470 pf ceramic capacitor, eliminating R20 and soldering a jumper wire across R54, while applying a 1N4148 diode across the potentiometer with the anode connected to the R54 jumper wire and cathode to ground, and replacing R21 by 4.7 Kohm 1% 1/4 watt resistor. After that modification, the current limit circuit will work like in the original UHU circuit.

On the next revision I will use a different (albeit more complex) way to take advantage of the intermittent current torque reserve of the motor.

Thanks,

Kreutz.

[kreutz]

I opened a Poll in order to receive feedback on what you want to see on a High Power UHU drive. The thread is: http://www.cnczone.com/forums/showthread.php?t=56037

[chrugel]

Hi
I just bougt a set of UHU controllers and one of these is supposed to drive a motor rated 130V 1A.
Can i just replace the IRF540 with IRF640 Fets since I'm not talking very high currents? Of cause all capacitors would also nee to be replaced with higher voltage rating.

Regards
Chris