Suggestions on variable DC power supplies/oscilloscope discussion.

[Cyparagon]

Re: Suggestions on variable DC power supplies.

a 50W switch-mode can be as small as 3 x 5 x 1”.

Good luck fitting a decent user interface on a surface that small

Another is size, how big is your workbench?

I'd recommend never sacrifice performance for the sake of space. Get a bigger workbench if needed.

On account of weight a 1000W Linear will need two guys, a hand truck, and some back-braces to install.

Eh, not really. HP6274B for example is 48 pounds, which is heavy, but most people could lift that without a problem.

there's the transient response of the power supply to consider

As you mentioned, this only matters in cases where there is a sudden and drastic change in load. Laser diodes don't ever do this, so it's not an issue.

If your oscilloscope uses Earth as the ground reference, you might get really bizarre readings.

If the power supply is floating (and the vast majority are), you need to connect the scope ground lead to one of the outputs anyway. Otherwise you're effectively measuring a voltage with only one terminal - it doesn't work. Of course you'll get bizarre readings if you don't measure things properly.

 

[Benm]

Re: Suggestions on variable DC power supplies.

Another option would be to look for a hybrid power supply. That basically works by providing a course supply from a compact switchmode unit, followed by a linear regulator that only needs to drop a couple of volts.

Adding output capacitors across the output terminals of a power supply, switchmode or linear, is NOT a good idea, since it completely defeats the effect of current limiting in the short term.

If you do this, you can only power a laser diode safely by turning on the power supply with the voltage set to zero and slowly increasing it until current limiting kicks in. If you were to set it to say 10 volts, limited to 200 mA, connecting a laser diode that runs at 200 mA at 4 or 5 volts would probably kill that instantly.

With any power supply it is wise to watch their turn-on behaviour on a scope. Even fully linear ones can have a spike on powerup due to delay in feedback circuitry.

For most uses this is not a problem, but if you're directly powering a laser diode off a lab supply it will be. Most of them were not designed with this application in mind: Most lab supplies are designed with constant voltage operation as the normal mode, and current limiting operation to protect connected circuits in case of mis-wiring and such.

They can overshoot output current briefly by quite a bit at the microsecond scale. For most devices this is no problem at all, but laser diodes can be fried within that timescale if unlucky.

 

[Illuminum3415]

Re: Suggestions on variable DC power supplies.

Cyparagon,

Thanks for the corrections, I knew we could count on you

Can you tell I don't work with lasers very much?

 

[Cyparagon]

Re: Suggestions on variable DC power supplies.

Most of them were not designed with [constant current] in mind

Maybe the cheaper chinese ones, but we shouldn't call chinese garbage the norm. If that were the case as you say, there'd be no need for a current control pot.

Another option would be to look for a hybrid power supply. That basically works by providing a course supply from a compact switchmode unit, followed by a linear regulator that only needs to drop a couple of volts.

You might not think so, but look it up some time: the stability/ripple specs on those are quite a bit worse than pure linear.

Even fully linear ones can have a spike on powerup due to delay in feedback circuitry.

Sadly, yes. Even moderately expensive ones can act like this. Current spikes are more common, where a PSU regulates voltage fine, but current mode is unstable on power-up.

Adding output capacitors across the output terminals of a power supply, switchmode or linear, is NOT a good idea, since it completely defeats the effect of current limiting in the short term.

But when not in current mode, it only serves to stabilize the voltage. Having high-ish current buck converters/drivers as a load for example can be problematic with low output capacitance. The supply will end up current limiting on the peaks of current draw even though average current may be well-within the rating of the supply. This causes problems for the buck converter. However, everything stabilizes if you add a big storage cap across the load.

 

[Down with Umbrella]

Re: Suggestions on variable DC power supplies.

Thank you everyone, I have learned so much about variable DC power supplies just in this thread. I will probably go with a switching ps. Desk space is not that much of a concern. I have room or can make rooms.
Cyparagon, you wouldn't have any O-scopes you are willing to part with do you? This winter is going to have a lull in work load so I was planning on a SSTC. Right now however I am doing 12s ugh.
Ps cyp. my iPhone couldn't load the video. I'll try my laptop tonight.

 

[Cyparagon]

Re: Suggestions on variable DC power supplies.

Scopes aren't something I have spares of. Not unless you want to buy my spare mostly-working 1950s 1MHz single-channel Tube-based scope

 

[DashApple]

Re: Suggestions on variable DC power supplies.

If your doing basic SSTC stuff I'd go for a rigol scope or similar

I have a dual channel 100mhz one and it dose well I think

 

[Down with Umbrella]

Re: Suggestions on variable DC power supplies.

Scopes aren't something I have spares of. Not unless you want to buy my spare mostly-working 1950s 1MHz single-channel Tube-based scope

Sounds like and old analog dinosaur. Ha

If your doing basic SSTC stuff I'd go for a rigol scope or similar

I have a dual channel 100mhz one and it dose well I think

Hmm those things aren't cheap. I see them listed for $3oo +
I'll keep an Eye out. Not necessarily in a rush to buy a scope but it's on my list.

Are tektronics poor quality machines ?

 

[Sigurthr]

Re: Suggestions on variable DC power supplies.

Tek's are awesome, but also awesomely expensive. Best bang for the buck is still the Rigol DS1052E (or some model similar, it's ~$400 new).

Under this price point you have to look for used, and sometimes into analog used. You need 2 channels with a >30MHz rating bare minimum for SSTC work, anything beyond that just improves quality of life exponentially. If you want to do DRSSTCs you'll need a DSO (digital storage oscope) or ASO (analog storage oscope), as a CRO (cathode ray Oscope) or DLO (digital live oscope) won't show burst data, only steady-state, continuous, and constant wave signals. That Rigol can do any type of TC work you throw at it and much more. I'm still using a 1970's 40MHz CRO, which is why you don't see any DR coils out of me.

 

[Down with Umbrella]

Re: Suggestions on variable DC power supplies.

Thanks for the suggestions sigurthr, how do Hanteks compare to the two listed above?

Realistically I see myself building a SSTC but not DRSSTC.

 

[Benm]

Re: Suggestions on variable DC power supplies.

Maybe the cheaper chinese ones, but we shouldn't call chinese garbage the norm. If that were the case as you say, there'd be no need for a current control pot.

Sadly it has become the norm on equipment priced at hobby-friendly levels. I'm sure the chinese standards play some role in this. If you buy something for a few hundred dollars you should realistically count on the current regulation not to be all that responsive.

For many applications this isn't a big problem since people often actually use it like an adjustable fuse, only requiring to kick in when they mess something up that causes a short.

But when not in current mode, it only serves to stabilize the voltage. Having high-ish current buck converters/drivers as a load for example can be problematic with low output capacitance. The supply will end up current limiting on the peaks of current draw even though average current may be well-within the rating of the supply. This causes problems for the buck converter. However, everything stabilizes if you add a big storage cap across the load.

In such cases i think you should add the capacitor after the output of the power supply. This would allow the supply to work in spec (i.e. at no moment exceed the set output current, however small the current draw peak) while allowing the buck converter to draw excess current from the added buffer cap.

Having output caps on the terminals can cause problems with laser diodes: if i set a power supply to maximum voltage but limited current, say limiting 30V/200 mA, a bunch of output caps would instantly fry any laser diode i connect to that power supply after it has been powered up.

The circuitry in the power supply has no control over current flow between an output cap at the terminals, it's just blunt capacitor ESR versus the laser diode.

A good current regulated power supply will not have output capacitors across the terminals - perhaps just a couple of low-esr nF range ones to help reduce any RF noise while not storing enough energy to blow something up.

 

[Cyparagon]

Re: Suggestions on variable DC power supplies.

Yes, when I said "across the load", I meant across the buck converter input, which is acting as the load in that respect.

Of course you don't use a big cap in current mode. Notice I prefaced the whole paragraph with "But when not in current mode"

 

[Sigurthr]

Re: Suggestions on variable DC power supplies.

Thanks for the suggestions sigurthr, how do Hanteks compare to the two listed above?

Realistically I see myself building a SSTC but not DRSSTC.

Sub-par compared to those two; in that price bracket a well cared for used Tek or new Rigol are the bee's knees, as they say. However, if you can find a good deal, really any well-cared-for fully operational scope is worth it as long as it meets min specs. Did I mention I'm using a ~45y/o clunker that only cost me $100? Hehe.

 

[ARG]

Re: Suggestions on variable DC power supplies.

Tek's are awesome, but also awesomely expensive. Best bang for the buck is still the Rigol DS1052E (or some model similar, it's ~$400 new).

There's a new entry level one from Rigol as well now, the 1054Z. Only slightly more ($50 or so IIRC) than the 1052E, and has four channels, larger screen, improved firmware ect. Amazing value for an entry level scope.

 

[Down with Umbrella]

Re: Suggestions on variable DC power supplies.

The rigol 1054Z is on Amazon right now for $399. Since I'm a prime member I would at least get two day shipping.
I would probably go with the upgraded Model. My cousin is in network communications and security business, I mentioned it to him and we agreed we could both benefit from owning one.

-------

Update: I bought the PS off Cyparagon.
Payment sent.

 

[Benm]

Re: Suggestions on variable DC power supplies.

With scopes on a budget you can go either way: Buy something old and analog but of good quality for a decent price ($100 will probably get you a long way for a good brand 20 MHz+ dual channel analog), or look at the digital ones such as the Rigol models instead.

There are ups and downs to either type.

Obvously the old analog/crt scopes are bulky, though usually quite luggable. They are usually less noisy to look at, and pretty good for continous signals. Often they can also cope with larger input signals if that were to happen by accident.

The new affordable digital ones are much more compact, and offer single shot trigger which is -extremely- useful when looking at the power-up behaviour of a power supply or driver.

If you have no scope at all buying an old second hand analog for as low as $50 would be a very good investment.

For a new good digital scope you're probably looking at $400+, and it may be worthwhile to get the used analog first and save up for the new digital. You will enjoy having both really, and you can use them together if you want to look at many signals at one time: many have trigger inputs and outputs, so you can trigger the timing of one scope from that off another.

This way you can combine two scopes if you need the extra traces. It's a bit confusing to work with, but certainly doable if you just need that extra channel or two and your signals aren't suitable for logic level analysers.