PDA

View Full Version : Cooling the Drives - Heatsinking and airflow


soulvoid
Fri 08 June 2007, 18:44
Why do you need an aluminum plate as the heatsink for the geckos? Wouldn't the steel back plate that everything mounts on work as a heatsink?

WTI
Fri 08 June 2007, 19:01
I would think you would want the heatsink to be on the exterior of the box so the heat is transfered outside and away from everything else.

If you look at a large audio power amp, the heatsink fins are on the outside.

Gerald D
Fri 08 June 2007, 22:55
Why do you need an aluminum plate as the heatsink for the geckos? Wouldn't the steel back plate that everything mounts on work as a heatsink?

That buries the Gecko's deep down into box (difficult for wiring, tuning). Besides, I wanted a fan to blow air onto the heatsink.

I would think you would want the heatsink to be on the exterior of the box so the heat is transfered outside and away from everything else.

If you look at a large audio power amp, the heatsink fins are on the outside.

Large power amps don't have a fan inside.

I am not fussy about how you lay out the control box. I wanted something I could build naked on a kitchen table, tune it, and then drop the working system into a box.

My approach is with a standard industrial enclosure, following normal industrial (factory) practice.

soulvoid
Sat 09 June 2007, 07:51
If I remember a discussion with some of my more audiophile friends (the kind that build amplifiers) Class A amplifiers are running at full power at all times, what does not go to the speakers creates heat so it's dumped on the heatsinks. Something like that.

Wasn't thinking about eventual wiring problems if the geckos are not lifted, will look at that. I think if you mount the geckos directly on the backplate circulating air behind the plate with a fan would be enough/essential. But in theory the 'heatsink' might as well be a steel plate as a alu plate?

Gerald D
Sat 09 June 2007, 09:42
Håvard, for the MechMate, we are not running the Geckos anywhere near their full capacity, and it is debatable whether we need heatsinks at all. Blowing a fan on them should be enough already. A good thermal mounting to the steel plate could also be enough. However, the method I employed (with alu + fan) should be good for Gecko's at full power and it also:
- brings the Gecko's to the front for easy wiring & tuning
- gives a nice alu grounding plate (particularly for the shields of the motor cables)
- shields the signal wires from the PMDX away from the high-power stuff
- saves space by "hiding" the fan under the Gecko's
- keeps the high-budget Gecko's nicely on a stage for display! :)

Out of curiousity, do you have any objections to the alu plate over the fan?

WTI
Sat 09 June 2007, 09:48
A copper plate would transfer heat even better than aluminum. Aluminum would transfer better than steel.

The cooler you can keep ALL of your electronic components, the longer they will last. Especially the capacitors. So if the Geckos get really hot (I have not run mine under load yet), I would want that heat to move out of the airtight enclosure and into the shop, rather than baking all the other components in the box with them.

Think of your computer as an example. The heat from the CPU is not sealed up inside the case just blowing around, it is blown outside of the box and cooler air is drawn in. We don't want to draw in dusty air...

The guys who build home theater PCs don't want to hear the CPU fans during the quiet parts of a movie. They mount huge, passive heat sinks (made of copper BTW) on the outside of their cases. No fans are needed, no noise is created.

My power amps have an A/AB switch on the back of them. If you set it at the AB setting, they run in Class "A" mode until they put out 35W then they switch to AB mode after that. If you leave them in "A" mode, they make great space heaters.

Commercial power amps used in concerts and nightclubs, have fans that kick on after they reach a certain internal temp. The front grills have replaceable air filters to keep the dust out. In a club, nobody can hear the fans run, and it saves space and weight by not having heatsinks on the exterior of the cases. The internal heatsinks are just aluminum foil, zig zaged like a car radiator.

A practical approach for us may be to mount the Geckos on the roof of our enclosures. Strip off the paint, and use some thermal compound to make a good thermal connection. On the outside of the enclosure, again strip off the paint and mount a heatsink fin array on the top. Dust every few months if they get too warm. Yes, we loose some efficiency going through the steel, but it is easy to make. If you wanted to get every BTU of use out of the heatsink, then you would have to cut the enclosure so that the Geckos mount directly to it.

Gerald D
Sat 09 June 2007, 10:10
A practical approach for us may be to mount the Geckos on the roof of our enclosures. Strip off the paint, and use some thermal compound to make a good thermal connection. On the outside of the enclosure, again strip off the paint and mount a heatsink fin array on the top. Dust every few months if they get too warm.

Horizontal surfaces are not good for convection cooling - you want naturally rising air to flow across the hot surface. Besides, a small layer of sawdust lying on top makes a good blanket.

I use the fan to dissipate ALL the hot spots inside the box and transfer the heat to ALL the box sides. (Don't forget that the major capacitors are NOT on the Gecko's) There is plenty of vertical surface area taking the front, back, and sides of the box together - more than a single heatsink on the back would give you.

Here (http://www.mechmate.com/forums/showthread.php?p=2676&postcount=9) I mentioned a test I did in my box - do you see anything wrong with that? :)

soulvoid
Sat 09 June 2007, 15:19
Out of curiousity, do you have any objections to the alu plate over the fan?

Not really... :) Just thinking K.I.S.S, why do I need a heatsink if the mounting plate will do the job? Secondly, I have trouble finding something to mount the plate with that looks neat, but that's really a no good reason

WTI
Sat 09 June 2007, 15:34
G,

How hot do the Geckos get in hard use? I know every system is different, but how about yours?

Gerald D
Sat 09 June 2007, 22:36
My Gecko's are set at 2 Amps (18K resistors). I havn't timed how long I can hold my hand on them, because my normal reaction is that they are cool enough not to keep holding on to them any longer. I could touch them until I get bored?

The G202's that I am using say in their manual: The G202 needs heatsinking for current settings greater than 6 amps. The case temperature (measured on the bottom plate) should not exceed 70 degrees C, and for best life should be kept to 50 degrees or less. Use heatsink compound between the G202 and the heatsink.

But, the preliminary manual on the G203's says:
The G203V needs heatsinking for current settings greater than 3 amps. The case temperature . . . . . (rest same as for G202)
I am a bit surprised that the current for the G203 has been halved, but this is probably because Mariss wants more people to use heatsinks, and not because the 203 produces more heat. The G203V actually produces less heat, as seen below:

The manuals also give the "power dissipation" of the drives:
G202 Power dissipation: 1 to 18 W (1 to 7 Amps)
G203V Power dissipation: 1 to 10 W (0 to 7 Amps)

I once asked (http://finance.groups.yahoo.com/group/geckodrive/message/11135) Mariss about the G201 and G202 family's heat and he said:

Model the drive as a 0.26 Ohm resistor. The 0.26 Ohm number is a
synthetic value that takes I squared R and switching losses into
account. Use the following equation which also includes quiescent
dissipation:

Watts = (I^2 * 0.26) + (0.015 * V)

Where I is the set phase current and V is the supply voltage.

The MOSFETs are rated at 33A continuous and have 0.04 Ohm 'on'
resistance. The current path is thru 2 MOSFETs and a 0.04 Ohm current
sense resistor. That makes for a 0.12 Ohm total resistance. The
difference, 0.14 Ohms, is the equvalent value for measured switching
losses. 13 - 14 Watts is still a lot of heat to get rid of from such a
small package.

In another post (http://finance.groups.yahoo.com/group/geckodrive/message/11208) he said about the G203V: The G203V can be modelled as a 0.22 Ohm resistor which gives 0.495W dissipation at 1.5A.

All those heat fins and things you see on audio equipment? Just realise that stuff gives off way more heat than the Geckos. We are talking less than 20 Watts here!

Here is another discussion (http://finance.groups.yahoo.com/group/geckodrive/message/11146) from that forum that is almost identical to how this thread started on this forum :)

domino11
Sun 09 September 2007, 18:40
Gerald,
Another good reason for extra heat sinking in electronics equipment is not always for the everyday running. The heat you generate when everything is ok and working properly is one thing, if a problem arises, then your dissipation in your drives could go up dramatically, say if a motor shorted or some other problem. Then the additional heatsinking may just save your drivers while you shut down the system. In my day job, I have done some work on power electronics designs and we always design the heat sinking for worst case scenarios, ie fault conditions and then add a little safety margin on top of that. So I heartily agree with your view on more is better. Also lower operating temperatures as others have mentioned means happier equipment. :)

Gerald D
Sun 09 September 2007, 23:32
During last autumn (fall?) ShopBot released a Gecko-based controller system, the "4g" (http://www.shopbottools.com/PRT_4g_Upgrade.htm) for driving their 1 amp motors. Apparently it mostly works, but this summer there have been reports of sudden shutdowns due to overheating. In that situation, a fan would help, heatsinks would help, placing the geckos further apart would help, orientating them vertical in the case would also help. Somehow I can't think of a more "wrong" way to mount those Geckos from a thermal viewpoint (unless a blanket is wrapped around them), yet they do manage to drive 1 amp motors (most of the time). The point of mentioning this is to illustrate the number of factors that must be taken into account when doing thermal design. And as Heath points out, you even need to design for when there is abnormal (fault) heat. The designer should consider abnormal to be normal. :)

But there does come a point when "enough" heatsinking/cooling has been added. Eg. for me it doesn't make sense to add external air and maintenance-hungry filters to drop the gecko's temp by a further 3 degrees. (The laws of diminishing returns . . . )

domino11
Mon 17 September 2007, 17:05
Gerald,
Good observation on the fans and filters. The most reliable electronic equipment only relies on convection cooling for the best performance. Fans and other mechanical ventiation systems will fail after a time and usually sooner in dirty or dusty environments. High speed bearings dont hold up well in these environments. Filters unless checked regularly will clog and reduce airflow. So if a fan fails or a filter gets clogged and your system fails thats not good. Proper heatsinking for worst case heat load is the best line of thought.

Gerald D
Mon 14 July 2008, 23:19
A good time to revisit this thread..... I read that folk are comparing the thickness of their alu plates when discussing heatsinks. This is putting the emphasis on a less important parameter.

The function of a heatsink fixed to a motor drive is not to absorb and hold the heat (then thickness would be important). There is a dual function to convey heat away and to dissipate the heat to the air around it. The folk talking of thickness, will do a good job of conveying, but are not addressing the dissipating.

To dissipate the heat to the air, we need two things:
- a large surface area for the aluminium to be in contact with the air, and
- a constant fresh supply of cooler air.

If one screwed a Gecko to a plain alu plate which is the same size as its base, then the surface area for cooling is hardly increased. In fact, you might be making it worse because the bond between the Gecko and the little plate could be poor. Making this little plate thicker, from say 1/8" to 1/4", will have practically no effect. But, if you make it 1/2" thick, and cut lots of grooves into it, to increase the surface area, then it becomes much more efficient. Simply making it 1/2" thick, without grooves", is hardly an improvement.

My approach has been to use a 3mm [1/8"] plate which is much bigger than the base of the Gecko, and to force air movement over that plate. I have been asked if 10mm [3/8"] will not be better . . . . It might be a little "better" but I seriously doubt if it will be significantly better. My gut feel is for 1 or 2 degrees lower temperature at the most.....even if the fan stops running.

3mm [1/8"] is enough for the plate to do its conveying function - making the perimeter edge thicker adds very little to surface area.

The question of thickness is similar to discussing wire gauge - if an 18 gauge wire is doing a reasonable job, why increase it to 10 gauge?

Richards
Tue 15 July 2008, 07:59
Gerald,

You're right. I use a large strip of 3/4-inch aluminum for my G202 drivers and 1/2-inch aluminum for my G203v drivers as a heatsink . The solid aluminum pulls the heat from the stepper driver, but it holds the heat (just like a thick frying pan spreads heat more evenly than a thin frying pan.) But, the point is, thick aluminum retains the heat.

In my application, where I normally run for about 1 to 2 hours at a time, the heat never builds much higher than room temperature, but if I were running all day long in a hot shop, I would use your method - 1/8-inch heatsink with a fan blowing on it.

javeria
Tue 06 January 2009, 10:03
Came across this for a heatsink - looks impressive at 4.8 USD.

http://www.dealextreme.com/details.dx/sku.13988

shelbygt500buddy
Tue 09 June 2009, 17:53
I know this post is a couple months old but i wanted to through something in... If you look at a computer cpu and look at the heat sink on that... some are copper most are aluminum you would me amazed how hot cpus run... If you are desiging a heat sink i would definitely design based on those... Also if you want the best performance you would wanna look into liquad cooling for computers you could get a unit and modify it for the box the liquads then now will not wreck electronics incase of leaks... you could also get a cooling system out of a old mini fridge etc and runn it through a vent to cool air to the heat sink.

riesvantwisk
Tue 10 November 2009, 10:04
I know this post is a couple months old but i wanted to through something in... If you look at a computer cpu and look at the heat sink on that... some are copper most are aluminum you would me amazed how hot cpus run... If you are desiging a heat sink i would definitely design based on those... Also if you want the best performance you would wanna look into liquad cooling for computers you could get a unit and modify it for the box the liquads then now will not wreck electronics incase of leaks... you could also get a cooling system out of a old mini fridge etc and runn it through a vent to cool air to the heat sink.

That sounds a bit overdone for a Mechmate :)

sailfl
Tue 10 November 2009, 16:04
Ries,

It is over kill for the Geckos. The aluminum plate works great.

riesvantwisk
Tue 10 November 2009, 17:49
Ries,

It is over kill for the Geckos. The aluminum plate works great.

Full ack :)
I was just adding this to my shopping list... it's time for me to by everything that runs current through...

Ries