Is powercycleing bad for test equipment?

[TheUnnamedNewbie]

Hi all. This is a topic that has been on my mind a lot lately. I seem to remember a few people saying that you should avoid unnecessary power-cycles of test equipment - specifically high precision ones. Form others I've heard the opposite - and keeping equipment on 24/7 will cost me more money, especially with the 30 eurocents/kWh that we pay around here.

The arguments for keeping them on 24/7 are generally along the lines of thermal cycles causing drift (I've mainly heard this in regards to anything with a very high precision oscillator). This makes sense on an intuitive level as well: every time you thermally cycle a device, the thermal gradients will cause different parts of your crystals and voltage references to expand and contract at different rates (as the thermal gradient passes along the device in question). This means that you are ever so slightly warping the device by doing so. This repeated warping will speed up the drift of the reference in question (be it frequency or voltage or whatever). Hence, you should try and keep references at a constant temperature at all times. (Some say this is also why a lot of instruments will keep the OCXO on whenever they are powered - but I've heard others say it's just so you don't have to wait half an hour before you can do your measurements)

The counter-argument I've heard is that a number of process that contribute to drift will increase in rate with temperature. Materials become more ductile causing them to flex more, and thus wear more quickly (in crystals). Diffusion rates increase, causing faster drift in semiconductor references or offset values. Additionally, a lot of other stuff just wears more quickly: Constant heat and load will decrease the lifespan of your electrolytic capacitors. VFD's lose their brightness. LCD burn-in is a thing. And in general - the longer something is on, in general, the more likely it is to fail.

So what is the right thing to do, and more important: Why? What are the different effects, and how important is each?

[bd139]

Turn it off if it's unattended. The first RIFA X class capacitor or schaffner filter that explodes will teach you that it is a fire risk. This is especially true as you have a lot of Philips kit. Every piece of Philips kit I've had has exploded.

[Gyro]

The first RIFA X class capacitor or schaffner filter that explodes will teach you that it is a fire risk. This is especially true as you have a lot of Philips kit. Every piece of Philips kit I've had has exploded.

Maybe it's time for you to get ahead of the game. 

[bd139]

Yeah I've got rid of it all

I have actually replaced them on sight but sometimes I can't be arsed

[Psi]

id be more worried about powering test gear off while it's still in the process of powering up.
Remember some test gear runs a full OS behind the scenes and you don't want your test gear to start giving BSOD.

Its entirely possible that there are some cpu cycles, out of the 50000000 it goes through at startup, which may 'brick' the gear if it happened to be halted right on them. 

[Fungus]

every time you thermally cycle a device, the thermal gradients will cause different parts of your crystals and voltage references to expand and contract at different rates (as the thermal gradient passes along the device in question). This means that you are ever so slightly warping the device by doing so. This repeated warping will speed up the drift of the reference in question (be it frequency or voltage or whatever). Hence, you should try and keep references at a constant temperature at all times.

Any decent piece of test gear will mount the references on a separate, tiny PCB or cut slots in the main PCB around the reference to prevent this.

[Cerebus]

every time you thermally cycle a device, the thermal gradients will cause different parts of your crystals and voltage references to expand and contract at different rates (as the thermal gradient passes along the device in question). This means that you are ever so slightly warping the device by doing so. This repeated warping will speed up the drift of the reference in question (be it frequency or voltage or whatever). Hence, you should try and keep references at a constant temperature at all times.

Any decent piece of test gear will mount the references on a separate, tiny PCB or cut slots around the reference in the main PCB to prevent this.

There are device stresses and board stresses. He's talking about device stresses, you're talking about board stresses.

As a general rule, don't worry about about power cycling kit unless it's precise enough to be used as a metrology transfer standard. So, yes keep your HP 3458A powered up 24/7 if you're lucky enough to have one and care about absolute accuracy as opposed to relative precision. Never power off a Fluke 731x 10V reference or similar device - they come with internal batteries to allow them to be 'hot shipped' for calibration and will light up the 'out of cal' light if they lose power. But turn off your HP 34401A and other similar lesser beasts.

However, one word of caution. The ageing and settling process can reverse in some voltage references and crystals if they are left powered off for too long. Don't leave equipment with highly precise internal references unused for a long time. If you aren't using it regularly leave it powered up for a few hours once a month (long enough for a 'full warm up', say 4 hours). This is also a good way to avoid the 'exploding RIFA capacitor syndrome' so well known to some of us.

[CatalinaWOW]

I don't think garden variety test equipment is any different than any other piece of electrical gear.  The most stressful thing that happens to well designed equipment is turning power on.  But there are risks to steady state operation also.  Particularly in switching power supplies the constant surge of current in filter capacitors heats and stresses them, resulting in capacitors being the number one failure point in older equipment.  Those same capacitors can act up if left off too long (usually years).

The upshot of this is that there isn't a general answer to this question.  For some gear life will be extended by powering it off, for others not. 

If I were looking for a general rule of thumb I would suggest turning off equipment overnight, but not between tests.  And exercising equipment every few months.  If you don't operate it that often, you probably don't need that equipment.