VNH5019 Thermal Issues above 20A

[Hans Dampf]

Posted on December 02, 2017 at 02:25

Hi,

I've got some problems in getting the max Current capability of 30A out of the Chip. As soon as the Current goes above 20A the Chip will sooner or later switch of Halfbridge B. I think the problem is to get the Heat fast enough out of the Chip on to the PCB, because the PCB and the Heatsink have only a Temperature of 50�C. According to the datasheet the Die of the Chip should have a Temperature of 150�C-175�C in order to switch off one of the Half-Bridges. 

So it seems that the Heat accumulates inside the Chip.

Therefore my question is if the HeatSlugs of the VNH5019 are capable of transferring the power, which is lost under full load, out of the Chip at all. Has someone ever achieved 25A or 30A?

By the way it doesn't matter if the applied Voltage is 12V or 24V.

The PCB Power traces from and to the Chip have a thickness of 70um on the TOP and the BOTTOM site.

The Heatsink touches the Copper area and the Top of the Chip.

The ambient Temperature is around 20�C - 22�C

[EDIT] There is no PWM frequency. PWM leads HIGH level all the Time. So there should be no switching losses.

[Electric Forest]

Posted on December 05, 2017 at 02:08

From 

http://www.st.com/en/evaluation-tools/eval-vnh5019-p1.html

  it seems like max continuous current is 12A.

Regards,

Lars

[Hans Dampf]

Posted on December 05, 2017 at 19:30

Yes but that's only for this particular carrier Board. It's limited because of the small copper Area and the fact that there is no Heatsink attached.

According to the datasheet the Maximum Continues Output Current is 30A

[Electric Forest]

Posted on December 05, 2017 at 22:12

I see. But 30A is still too much. I think the datasheet is a bit inaccurate when it says 'Output current: 30A' on the first page. If you look at section 2.1, the absolute maximum rating is 30A - this means that above this it can be damaged. So you should keep distance to 30A. Also, look at the Rds and calculate power dissipation. Note that Rds increase with temperature from the typical value.

I would still think 20A should be achievable (but a stretch) if you have very good thermal design. Do you have solder underneath the package to give a good thermal connection to the board? Do you have vias to the bottom side to lead heat that way?

Cheers,

Lars

[Nickname15513_O]

Posted on January 08, 2018 at 13:03

Hi Hans,

We confirm that VNH5019 max current capability is 30A, since this is the min current limitation. On the other hands, managing such high power inside the chip, you have to take under consideration also the thermal performance in terms of Rthj-case and Rthj-amb (as specified in the datasheet).Especially the Rthj-amb depends on the pcb layout and number of layers (please refer to Chapter 3 in the datasheet for data).

As soon as the silicon will reach the Thermal shutdown temperature the device will be protected and will shut-down.

In order to perform further trials on different load profile we suggest you to run some simulation on TwisterSIM tool: a unique electro-thermal simulator that helps shorten the design solution cycle by enabling complex engineering evaluations. TwisterSIM is available for free at

http://www.st.com/twistersim

Best regards

Davide

[Electric Forest]

Posted on January 10, 2018 at 00:50

This can't be correct. The absolute maximum rating is 30A. The device is not supposed to operate at this high current, the datasheet says: 'Stressing the device above the rating listed in the “absolute maximum ratings� table may cause permanent damage to the device. These are stress ratings only and operation of the device at these or any other conditions above those indicated in the operating sections of this specification is not implied. Exposure to absolute maximum rating conditions for extended periods may affect device reliability. Refer also to the STMicroelectronics SURE program and other relevant quality document.'

Another approach is to calculate power dissipation and die temperature. The datasheet specifies high side Rds,on is up to 26.5mOhm and low side Rds,on is up to 11.5mOhm at 15A. Assuming the Rds,on is the same at 30A (but it is most likely higher), we get a power dissipation of 34W at 30A. With the largest copper area, the datasheet says 20C/W thermal resistance for the high side and 15C/W for the low side (figure 20 and 21). Assuming the same temperature across the die, the effective thermal resistance will be 8.6C/W. Temperature difference from die to ambient will thus be 34W*8.6C/W = 292C. In reality, the high side will be even hotter because it dissipates more power and has a higher thermal resistance to ambient. At 20A the temperature difference will be 131C. At 15A it will be 74C.

Conclusion:

You can safely drive a motor with a continuous current up to somewhere between 15A and 20A if your thermal design is very good.

Note that all this is assuming no PWM as in the case of OP. If the duty cycle is kept below 50% you might reach 30A (but should never go too close to or above that level no matter how low duty cycle due to the abs. max rating).

Best Regards,

Lars Eirik

[Hans Dampf]

Posted on January 12, 2018 at 00:54

Hi,

Well first of all thank you all for your advise. I would have written earlyer but the hollidays catched all my attention and there seems to be also a problem with the Forum. If I'm not logged it says literally every time I tried to view this post that an unexpected Error has occured. Yeah Nevermind.

@

nicolosi.davidegaet

I was not sure if that is really the case. I've searched the whole Web in reference to the VNH5019,VNH2SP30,VNH3SP30 to find a Design which is capable of the full power. Most licely it whould have been the VNH5019 because of it's lower RdsOn but I never find some. The best I could find was the Carriers from Pololu which are capable of a continues 16A per Channel.

So I thought no one ever tried.

[Edit] If I have time I will take a look at you programm. Maybe that will show of some other issues.

@

lars.eirik.mobak

At some point in the design process I think I've calculated it but I don't know anymore. As we will later see this wasn't (not yet) an issue.

And now the most important news: I've made some progress :)

The key was into first place the Thermal Coupling between the Heatsink and the Top of the Chip. As caused by design the Heatsink is not able to sit directly on to the Chip. There is a 2mm thick copper plate between them and the Coupling through to the Heatsink was thery bad. So what I did was I simply added some

Thermal Compound between Chip->Copper Plate and Copper Plate->Heatsink.

Now the Thermal resistance is much lower. The Coupling is simply much more 'direct '.

At the first test on 20A it needed a lot of time to Warm the Heatsink up. At reached 85°C I've putten the Fans on, because of the high Temperature, but with plenty of reserves left.

The highest continues current I'm achiving now is 25A without switching off but thats really impractical because of the large Fan and the fact that the Temperature Delta between the Temperature of the Heatsink and the Point of Switching off is just 5°C.

At 24A it looks much better. I'm now able to use the small 25x25mm Fans on top of the Heatsink too cool it.

Never the less the Temperature Delta is still 5-10°C. At Ambient Temperatures of up to 25°C that's not a problem but I think in summer I'm getting some trubble.

So the best current I think I can savely run is 23A. At ambient Temp of 22°C and with the Fans on, the Heatsink gets to about 50-55°C with a delta of 15-20°C.

So there you have it. I think that's all I can get with the actuall revision of my Board. Maybe with some cooling improvements 24A but I think that's another revision of the Board ;)

Theres still some potential left.

The solder Pads on the PCB for the Chip doesn't have Vias yet for directly connecting the Bottom to the Top Layer. They are all allocated around the Chip in the Power Traces.

[Electric Forest]

Posted on January 12, 2018 at 01:28

A fan definitely helps a bit! It will also help to blow at the PCB itself. And the secondary side.

If you can somehow remove the copper plate that will help also.

When you rev the board, if you have space, I think it would be better to use two devices in parallel.

Cheers,

Lars