Test Equipment Anonymous (TEA) group therapy thread

[tonyalbus]

Testing some of the OCXO TCXO OCVCXO etc... stuff from Ali....
Tested and found the the connections, how much power the draw, on start-up on idle.. temp...
And the specs for you in PDF...

CTi OSC5A2B02
CTi OC5SC25
Vectron C4550A1-0213
Trimble 34310-T

[Cerebus]

Heads up for the USAians: Relatively cheap Keithley branded tri-lug Triax leads: https://www.ebay.co.uk/itm/133530762264

Standard three lug Triax to ultra miniature Triax, 6 foot, £39.25 each in proper money.

[Cerebus]

Enjoying my PCBGrip this weekend.
It is immensely cool.
So far the only Kickstarter that made sense for me.
 

That actually looks quite useful.

[Kosmic]

mnem
 

Look like output impedance is fix at 50 . On the Rigol you can set that manually and it will automatically adjust amplitude based on selected output impedance. I often use that feature.

Mmmhmmm... that's not a feature available on every Sig gen tho. I wonder if that tradeoff was necessary to get the improved output level linearity?

@tautech - what do you have that's comparable, and what gotchas do you know about on this unit?

EDIT: Manual is here: http://68.168.132.244/PSG9080_EN_manual.pdf

mnem
 

who says linearity is better on the Juntek than Rigol ? you found a comparison somewhere ?

[Kosmic]

Yeah, not since they showed up on Dave's bench; typical now is $350-400. Now on CLEARANCE for $300 +S/H at TEquip. From the video this one appears to have much more linear output re: voltage vs frequency. Rigol still has yet to address this with FW even in the more expensive models, to my knowledge. Plus this one is rated 300MS/s vs 125MS/s for the Rigol.

mnem
 

PSG9080 at 163$ US on Ali: https://www.aliexpress.com/item/1005001507207074.html?spm=a2g0o.productlist.0.0.20bc53b9tnKOqc&algo_pvid=624891f7-d70e-4a59-8347-265f64056dc6&algo_expid=624891f7-d70e-4a59-8347-265f64056dc6-0&btsid=0bb0624616012220981132182e5bf8&ws_ab_test=searchweb0_0,searchweb201602_,searchweb201603_

I guess it's not a bad price.

Look like output impedance is fix at 50 . On the Rigol you can set that manually and it will automatically adjust amplitude based on selected output impedance. I often use that feature.

Also the ground on the output connectors is floating, not referenced to ground.

[bd139]

Don’t do the junktek. It’s a piece of trash. Check the tear downs. I know it’s slightly more expensive but the ass end Rigol and Siglent units are better value for money.

To clarify the impedance and amplitude thing it’s pretty lame implementation usually. There is no ALC feedback or anything like that which you’d expect in a decent signal generator. Basically source is 50 ohms. Into a high Z load that’s insignificant so amplitude is set against a peak reading voltmeter for a bunch of frequencies and the cal curve is build out if that to control the amplitude. 50 ohms, well you just half the amplitude from the cal curve. Job done! 50 ohms source and load is a 50% voltage divider. Then if you know the load impedance you can calibrate it in dBm doing a simple calculation.

Really I want a nice signal generator that has a wide band source, ALC and then a stepped attenuator rather than this arrangement but it’ll do for now. (Siglent that is). I might build one. They aren’t that difficult to put together.

[mansaxel]

Eish that's a nasty job. Hope the switch is repairable!

I hope so. The bushing that holds the switch decks to the front panel has been pushed off the deck. It's a press fit originally, and if the axles are getting a beating from the front this might happen. If I can't press it together, I think I'll go for gluing it.

[bd139]

That hurts to look at 

[Kosmic]

Don’t do the junktek. It’s a piece of trash. Check the tear downs. I know it’s slightly more expensive but the ass end Rigol and Siglent units are better value for money.

+1

I think the same.

[tautech]

Don’t do the junktek. It’s a piece of trash. Check the tear downs. I know it’s slightly more expensive but the ass end Rigol and Siglent units are better value for money.

The first thing I noticed in Scott's (Defpom) teardown was the SMPS powering it. 
Really, who wants one of those powering your signal source ? 

To clarify the impedance and amplitude thing it’s pretty lame implementation usually. There is no ALC feedback or anything like that which you’d expect in a decent signal generator. Basically source is 50 ohms. Into a high Z load that’s insignificant so amplitude is set against a peak reading voltmeter for a bunch of frequencies and the cal curve is build out if that to control the amplitude. 50 ohms, well you just half the amplitude from the cal curve. Job done! 50 ohms source and load is a 50% voltage divider. Then if you know the load impedance you can calibrate it in dBm doing a simple calculation.

When the output is set to 50 ohms impedance dBm units can be selected in Siglent AWG's.
Enter the value required and the unit selection pops up, select dBm !

[bd139]

Yeah aware of that. The unit does the calculation for you 

[Saskia]

just returned from a haul and missed out on a Canon Extender 2x Mark 3.
Bloody fucking hell.

On the plus side:
2 Gossen Metra Hit 28S for 100 Pesos apiece
1 Beckman Industrial SI-9000 differential probe 25MHz 50 Pesos
1 Extech 382222 Lab Power supply 10 Pesos
12 boxes with content (unsorted and uncataloged) 60 Pesos.

If I saw correctly there is a shitload of coils and inductors (for SMPS), some filter capacitors, lots of SMD components, MOSFETs, ... in those boxes.

Will have to have a closer look.

[mansaxel]

On the plus side:
2 Gossen Metra Hit 28S for 100 Pesos apiece

Gossen fan here. Seriously envious! That's borderline jammy git.

[Saskia]

btw why do you need a bench DMM if your handheld already does 5.3/4 digits ....

[Saskia]

On the plus side:
2 Gossen Metra Hit 28S for 100 Pesos apiece

Gossen fan here. Seriously envious! That's borderline jammy git.

Sorry, one of them is already going to a forum member who asked to get one for him and pointed me towards the seller. Turns out that he was the lead developer auf Braun who back in the 80s did some seriously good hifi stuff.
He's 80 and slowly retiring and his daughter thought she should get rid of his stuff.

So I filled my Skoda with his drawer thingies that were full of new parts.
Those are the real grabs. But, I missed out on that Canon converter and that seriously annoys me.

[mnementh]

www.youtube.com/b3yFxg2LwXI

mnem
 

Look like output impedance is fix at 50 . On the Rigol you can set that manually and it will automatically adjust amplitude based on selected output impedance. I often use that feature.

Mmmhmmm... that's not a feature available on every Sig gen tho. I wonder if that tradeoff was necessary to get the improved output level linearity?

@tautech - what do you have that's comparable, and what gotchas do you know about on this unit?

EDIT: Manual is here: http://68.168.132.244/PSG9080_EN_manual.pdf

mnem
 

who says linearity is better on the Juntek than Rigol ? you found a comparison somewhere ?

To get the BW this one advertises on the bottom-of-the-line Rigols and Siglents, you have to hack them. Then output linearity goes totally out to lunch when you get up in those frequencies; its a constant topic of conversation in the hack/mod threads. Of course, since they aren't meant to operate up there, not the MFR's fault.

Don’t do the junktek. It’s a piece of trash. Check the tear downs. I know it’s slightly more expensive but the ass end Rigol and Siglent units are better value for money.

The first thing I noticed in Scott's (Defpom) teardown was the SMPS powering it. 
Really, who wants one of those powering your signal source ? 

To clarify the impedance and amplitude thing it’s pretty lame implementation usually. There is no ALC feedback or anything like that which you’d expect in a decent signal generator. Basically source is 50 ohms. Into a high Z load that’s insignificant so amplitude is set against a peak reading voltmeter for a bunch of frequencies and the cal curve is build out if that to control the amplitude. 50 ohms, well you just half the amplitude from the cal curve. Job done! 50 ohms source and load is a 50% voltage divider. Then if you know the load impedance you can calibrate it in dBm doing a simple calculation.

When the output is set to 50 ohms impedance dBm units can be selected in Siglent AWG's.
Enter the value required and the unit selection pops up, select dBm !

@tautech - Oh, come on... SMPS are in everything. I suppose I should chunk all my 2465s because of that too.     Now if it's a bad implementation, sure. But that's true of every bit of test gear ever made.

@bd - Thanks, that's the kind of insight I was looking for. I DID look at the teardowns. This looked like a lot of bang/buck; I was asking for constructive response, not just "It's junk". I expect it to be cheaply made; it's like the next generation of the FeelTech ARB for duck's sake. I'm not seeing anything here that I didn't see there, and the FeelTech costs only a few $$ less for much lower resolution and again, really horrible output level linearity.

Bottom line is I'm not going to spend $300-400 on a SG anytime soon. I expected it to be slightly better than the FeelTech; which really is all I need for my ESC tinkery. I figured a few bux more for a little better wasn't too bad.

For better absolute quality like bd139 was saying, I can get a WaveTek 145 for $50 from AllTest.

mnem
 

[Saskia]

@mansaxel fyi the BD232 adapter for the metra hit was included in the haul.
(for each of the DMMs ...)

[tautech]

@tautech - Oh, come on... SMPS are in everything. I suppose I should chunk all my 2465s because of that too.     Now if it's a bad implementation, sure. But that's true of every bit of test gear ever made.

mnem
 

This ^^^ and how do we know if it's good or not ? 
Detailed analysis is required .....do you wanna do it ?

Another Yaigol thread in the making ? 

[mansaxel]

@mansaxel fyi the BD232 adapter for the metra hit was included in the haul.
(for each of the DMMs ...)

That is downright taunting! Keep at it! 

[bd139]

@tautech - Oh, come on... SMPS are in everything. I suppose I should chunk all my 2465s because of that too.     Now if it's a bad implementation, sure. But that's true of every bit of test gear ever made.

mnem
 

This ^^^ and how do we know if it's good or not ? 
Detailed analysis is required .....do you wanna do it ?

Another Yaigol thread in the making ? 

Actually it's fairly easy to evaluate it. The main SMPS is not bipolar and kicks out 5 volts. There are at least two boost converters on the board to generate the required voltage rails for the output amplifiers. That's almost guaranteed to raise the noise floor of this pretty high.

The other vendors, Rigol and Siglent, use a primary SMPS to generate multiple rails then have quite hefty filtering and linear regulation for the rails.

It's a box of cut corners that's what it is.

[Kosmic]

www.youtube.com/b3yFxg2LwXI

mnem
 

Look like output impedance is fix at 50 . On the Rigol you can set that manually and it will automatically adjust amplitude based on selected output impedance. I often use that feature.

Mmmhmmm... that's not a feature available on every Sig gen tho. I wonder if that tradeoff was necessary to get the improved output level linearity?

@tautech - what do you have that's comparable, and what gotchas do you know about on this unit?

EDIT: Manual is here: http://68.168.132.244/PSG9080_EN_manual.pdf

mnem
 

who says linearity is better on the Juntek than Rigol ? you found a comparison somewhere ?

To get the BW this one advertises on the bottom-of-the-line Rigols and Siglents, you have to hack them. Then output linearity goes totally out to lunch when you get up in those frequencies; its a constant topic of conversation in the hack/mod threads. Of course, since they aren't meant to operate up there, not the MFR's fault.

Just a small calibration problem

The hardware in the DG811 and DG992 is exactly the same.

[bd139]

It is, and the siglent. The cal instructions for the siglent give that away.  The 33120A does the same.

Edit: dang seen a reasonably priced decent condition HP 48G on ebay. Thinking about it. It's not a GX but it'll do 

Edit 2: chickened out. Need more RAM.

[tautech]

@tautech - Oh, come on... SMPS are in everything. I suppose I should chunk all my 2465s because of that too.     Now if it's a bad implementation, sure. But that's true of every bit of test gear ever made.

mnem
 

This ^^^ and how do we know if it's good or not ? 
Detailed analysis is required .....do you wanna do it ?

Another Yaigol thread in the making ? 

Actually it's fairly easy to evaluate it. The main SMPS is not bipolar and kicks out 5 volts. There are at least two boost converters on the board to generate the required voltage rails for the output amplifiers. That's almost guaranteed to raise the noise floor of this pretty high.

The other vendors, Rigol and Siglent, use a primary SMPS to generate multiple rails then have quite hefty filtering and linear regulation for the rails.

It's a box of cut corners that's what it is.

Yeah well they haven't even considered the RF that a SMPS produces can be shielded from the SG circuitry by a metal divider or even protect the lab from SMPS RF by installing the whole thing in a metal enclosure.
OK they made the shipping cheaper by omitting proper shielding but really is that of value for a signal source ? 

Anyways, it should make a fine noise source for the bench ! 

[bd139]

I think in the Junktek they're going on the fact that the box is mostly empty and you can put the SMPS miles away from the main board.

[Specmaster]

@bd - Thanks, that's the kind of insight I was looking for. I DID look at the teardowns. This looked like a lot of bang/buck; I was asking for constructive response, not just "It's junk". I expect it to be cheaply made; it's like the next generation of the FeelTech ARB for duck's sake. I'm not seeing anything here that I didn't see there, and the FeelTech costs only a few $$ less for much lower resolution and again, really horrible output level linearity.

Bottom line is I'm not going to spend $300-400 on a SG anytime soon. I expected it to be slightly better than the FeelTech; which really is all I need for my ESC tinkery. I figured a few bux more for a little better wasn't too bad.

For better absolute quality like bd139 was saying, I can get a WaveTek 145 for $50 from AllTest.

mnem
 

Exactly that, if you're serious stuff then something like a Siglent etc is justified, but for the sort of tinkering most hobbyists do, I see nothing wrong with that JunTek, or indeed the Feeltech units, cheap and cheerful, just like the TinySA etc. I have a DSO138 2.4" scope, which for rough and ready signal tracing in audio gear is perfectly OK and helped me on many occasions. Yes I have far better scopes at my disposal, but the DSO138 is just so quick to power up and connect to, it is often used in preference to show the area of concern on amplifiers for example.