Author Topic: EEVBLOG 512 Possible Rigol regulator overheat fix which works with over products  (Read 1410 times)

0 Members and 1 Guest are viewing this topic.

Offline IanTaigTopic starter

  • Newbie
  • Posts: 2
  • Country: au
Hi All,

Was just watching episode #512 about the overheating regulator in the Rigol 832 and I recall a similar problem with their DC loads as well. I have worked many years in the industrial electronics area and it is amazing how common this problem actually is and not just with regulators but many electronic components.

The problem I have found mainly from the point of view that electronic products are mainly designed for either European, US or Japanese environments, the closest to us being European with a 220 to 230V nominal supply which is usually 240V or less at maximum. Australia on the otherhand is usually closer to 250V especially now with the prevailance of solar. For example in the lab at work I get about 247V on 2 phases and 249V on the 3rd. This is within spec, here at home it is about 245V normally but when the solar is going full tilt in the summer it gets upto 250V. I have also worked on equipment from mining areas over in WA and some areas out there can get up to 265V. Anyhow long story short, the voltage is pretty high.

The problem with this excessive voltage is it does put a lot of stress on some components, in the case of some Rigol products with linear power supplies this is excessive heat in the regulators even though in the case of the DP832 the heatsink is also inadequate. Other common problems I've seen with equipment is rapid heater element aging, burned out encapsulated transformers, premature electrolytic capacitor failure and bootstrap/voltage balancing resistors in power supplies going open, oh and on older equipment, rapidly burning out indicator lamps.

So how I get around this problem is actually quite simple, non invasive and considerably lengthens the life of many products. What I do is use a 300VA toroid transformer (or any transformer with the same ratings will do) with a 30V secondary, I connect this as an auto-transformer so I get about 220V coming out at the tap connection (Provided the transformer windings are connected with the correct phasing). This setup is good for upto 2400VA of load, I pretty much run everything in the lab through a couple of these, I also have a 3 phase unit too but I generally use that to test our products at lower voltages (385V @ 40A) that we export to Europe. I also make it a point to ensure that Neutral is common to both the mains and load sides of the auto-transformer and I also have a 2 Pole 10A C curve circuit breaker on the load side. Really important, check the output voltage before you use this, if the voltage is higher than the incoming mains then the transformer is incorrectly phased and you will need to either swap the primary leads or the secondary leads but not both.

I'm not sure what the mains voltage is where Dave is but assuming it is around 245V then dropping the voltage down to about 220V should reduce the dissipation in that LM317 by about 20% so the heatsink will still be way too hot IMO but that and the junction temperature will have come down by quite a bit. This should also reduce the dissipation in the series pass elements too as the DC bus voltage for each channel will have come down too therefore the series pass element doesn't need to dissipate quite as much.

I don't have a DP832 but plan on buying one in the next week or 2 so I will try it with and without the auto-transformer and see what the difference is and will make a further post then. If anyone else wants to try this out you can also test it using a variac (which is a variable auto-transformer) and just wind the voltage down to 220V and see the result.
 

Online 2N3055

  • Super Contributor
  • ***
  • Posts: 6454
  • Country: hr
Hi All,

Was just watching episode #512 about the overheating regulator in the Rigol 832 and I recall a similar problem with their DC loads as well. I have worked many years in the industrial electronics area and it is amazing how common this problem actually is and not just with regulators but many electronic components.

The problem I have found mainly from the point of view that electronic products are mainly designed for either European, US or Japanese environments, the closest to us being European with a 220 to 230V nominal supply which is usually 240V or less at maximum. Australia on the otherhand is usually closer to 250V especially now with the prevailance of solar. For example in the lab at work I get about 247V on 2 phases and 249V on the 3rd. This is within spec, here at home it is about 245V normally but when the solar is going full tilt in the summer it gets upto 250V. I have also worked on equipment from mining areas over in WA and some areas out there can get up to 265V. Anyhow long story short, the voltage is pretty high.

The problem with this excessive voltage is it does put a lot of stress on some components, in the case of some Rigol products with linear power supplies this is excessive heat in the regulators even though in the case of the DP832 the heatsink is also inadequate. Other common problems I've seen with equipment is rapid heater element aging, burned out encapsulated transformers, premature electrolytic capacitor failure and bootstrap/voltage balancing resistors in power supplies going open, oh and on older equipment, rapidly burning out indicator lamps.

So how I get around this problem is actually quite simple, non invasive and considerably lengthens the life of many products. What I do is use a 300VA toroid transformer (or any transformer with the same ratings will do) with a 30V secondary, I connect this as an auto-transformer so I get about 220V coming out at the tap connection (Provided the transformer windings are connected with the correct phasing). This setup is good for upto 2400VA of load, I pretty much run everything in the lab through a couple of these, I also have a 3 phase unit too but I generally use that to test our products at lower voltages (385V @ 40A) that we export to Europe. I also make it a point to ensure that Neutral is common to both the mains and load sides of the auto-transformer and I also have a 2 Pole 10A C curve circuit breaker on the load side. Really important, check the output voltage before you use this, if the voltage is higher than the incoming mains then the transformer is incorrectly phased and you will need to either swap the primary leads or the secondary leads but not both.

I'm not sure what the mains voltage is where Dave is but assuming it is around 245V then dropping the voltage down to about 220V should reduce the dissipation in that LM317 by about 20% so the heatsink will still be way too hot IMO but that and the junction temperature will have come down by quite a bit. This should also reduce the dissipation in the series pass elements too as the DC bus voltage for each channel will have come down too therefore the series pass element doesn't need to dissipate quite as much.

I don't have a DP832 but plan on buying one in the next week or 2 so I will try it with and without the auto-transformer and see what the difference is and will make a further post then. If anyone else wants to try this out you can also test it using a variac (which is a variable auto-transformer) and just wind the voltage down to 220V and see the result.

Problem with watching old videos is that many times it doesn't apply anymore.
That problem with overheating was solved years ago..

Also that applies to reported software problems with firmware that is 10 levels downlevel from what is today..
And I mean that for all devices, not just this one...

 

Offline IanTaigTopic starter

  • Newbie
  • Posts: 2
  • Country: au
Nice to know it has been fixed. The higher voltage here can and does present quite a problem out in the field, particularly with linear supplies as they do end up dissipating a lot more power. In industrial electronics it becomes even more of an issue because the ambient operating temperatures for this equipment is generally higher here in the summer time compared to places like Europe. The temperature difference is quite a lot just by dropping the voltage a bit.

Switch mode supplies generally aren't too much of a problem, they are usually designed to operate 90-265V so should in theory have adequately rated components although I have found myself replacing 400V electrolytics with a 450V device. Voltage stress on resistors is another common problem, they tend to go higher in resistance as they age until they go open completely, this I usually fix by replace a resistor with 2 in series. Other common problems I've found are cooked snubber networks, VDS breakdown on MOSFETs in flyback power suppies. The list goes on and on.

Most problems are long term and usually happen just after warranty runs out. I just make it habit to drop the voltage down because Murphy always makes it a habit to cause something to fail when you need it most.

As a general rule of thumb, a semiconductor will have is life halved for every 10 kelvin rise in temperature and from a lot of the product validation testing we've done this seems to be a pretty good approximation. One of the tests we do for IEC certification actually requires us to run our product at 6% above the maximum supply voltage rating and measure the temperature of all the components, that 6% may sound insignificant but it makes a huge difference on the operating temperature of the components (generally a 10-15% increase), most components are not allowed to have a temperature rise exceeding 50 kelvin above ambient under those conditions.

Believe me, having to design equipment that has some pretty strict guildlines, you become very much aware of the problems external influences have on a piece of equipment, even influences most people wouldn't even think about (Ever wondered why some products have an altitude specification?).

I must say something about that Rigol 832 video, it certainly does demonstrate the robustness of the LM317 voltage regulator, they've been around forever and are just so damn reliable.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf