I am afraid not, different voltages on positive and negative rails could not be set it is only symmetrical setting.
The VQFN smt package TPS7Axxxx regulator chip is tricky to solder, no exposed legs and a belly pad. But like Passive420 mentioned, it can be done with a hotplate and hot air method. If you haven’t had the pleasure to solder these successfully, you better get extra’s for practice 😂.
I used the Chinese 'sigma 79' and 'sigma78' tpsa based 3 pin regs on my PSU1. Unfortunately not available any more. They don't like/need much capacitance after them so I just put 47uF onboard Miro's PCB on the outputs. The dac still has all the onboard local decoupling. Works a treat. They're are other implementations out there. Interested in Brijacs project.
I came to these sigmas after first using one....just one in a Cambridge Audio Dacmagic. It went from so so after lots of tweaks to being a very nice sounding DAC.
The tpsa have very very low output impedance. Probably the reason they sound good.
Here before I swapped out the big output caps.
https://www.diyaudio.com/community/...st-tht-i2s-input-nos-r-2r.354078/post-6677715
I came to these sigmas after first using one....just one in a Cambridge Audio Dacmagic. It went from so so after lots of tweaks to being a very nice sounding DAC.
The tpsa have very very low output impedance. Probably the reason they sound good.
Here before I swapped out the big output caps.
https://www.diyaudio.com/community/...st-tht-i2s-input-nos-r-2r.354078/post-6677715
Quite small, but can be soldered by hand if you have solid experience in smd soldering. In my designs belly is always TH pad, so it can be soldered from below. I did one to test, but i have directly heated tip soldering station, and a choice of tips. But yeah, i use hotplate station, much easier and less chance to destroy the chip.
Here you can see, gnd plane is between input and output part of the regulation circuit, which is important. Also impedance of the circuit, LC filter can be formed with resonant frequency that can cause oscillation. Depending on the layout and caps used, care must be taken.
Here you can see, gnd plane is between input and output part of the regulation circuit, which is important. Also impedance of the circuit, LC filter can be formed with resonant frequency that can cause oscillation. Depending on the layout and caps used, care must be taken.
flux is really your best friend, I am now going to the pub with it ! It makes really your life simplier and easy and way cooler ! Yup it is really feasible to solder this package with a 0.5 tip. One thing though : all those flux that spreads below the parts as well as a bigger soic is hard to remove even with isopropyl ! And that is a problem for stray capacitance...
In this case, and for our use cases, not that much of an issue. RF gear and precision amplifiers (and even there, layout is waaaaay more important), yes, and those also have issue with flux residues that were not cleaned on top of the board, due to surface leakage (order of magnitude several hundred pA usually, depending on flux quality, residue thickness and size). For this, not so much, just use solid quality flux. If i may suggest, subscribe to altium mailing list, very nice material and they tend to invite you to online seminars which are amazing, lectures from many well known engineers regarding circuits and pcb designs, caveats etc.
Not sure with the very fast op amps....
Altium curses, Yup , i am a fan, there is guy that always seems under coco...very funny.
Altium curses, Yup , i am a fan, there is guy that always seems under coco...very funny.
not sure it changes anything when you see skin oil is enough when not cleaned on a pcb. Now it is about to know if it changes anything about the signal intigrity related to our ears....
I use a 310 M hz flat bandwith at unity gain op amp on my 4 layers board, so I do not want to risk. But I have seen when removing an op amp some flux spreaded below whatever the cleaning ! Now for several reason not only flux related, I tend to bon a little the soic op amps pin to have a space high enough between the soic bottom and the pcb for cleanoing purpose (means your op amps have not a gnd bullly that needs to be soldered on a central pad fpr Tempco cooling) ... It is not scientifical as nothing is measured, just seemed logic related to all the papers related to high dpeed and tp layer leakage (stray capacitance spread) .
Makes me sleep better and reduce my french wine bottles/week ratio ! (the reactor here is myself)
Last edited:
Yes, if you need so high current.
For lower current we have TPS7A4901/3001, or better one TPS7A3901.
10-20 sec, if you have hot air gun. 🙂
At least like this: https://www.aliexpress.com/item/4000085968033.html?spm=a2g0o.productlist.main.3.3fb9DMHaDMHa3n&algo_pvid=c4665ac6-95b6-4823-9f82-b8190ad01749&algo_exp_id=c4665ac6-95b6-4823-9f82-b8190ad01749-1&pdp_npi=4@dis!ILS!158.56!95.15!!!42.01!25.21!@2102f67717238790965627739e8308!10000000229441543!sea!IL!137829664!X&curPageLogUid=2L3l4TlvVKEg&utparam-url=scene:search|query_from:
Even w/o hotplate.
Try spending $25 and you'll prefer any smt (maybe other than BGA) over TH.
Alex.
Last edited:
Is everyone here a professional electronics engineer? You guys talk things which way over my knowledge level.
I've got a little old scope and desoldering station recently so am slowly embedding myself in this hobby (after doing stuff for 10+ years of on and off ).
I've got a little old scope and desoldering station recently so am slowly embedding myself in this hobby (after doing stuff for 10+ years of on and off ).
Smaller yes, better no. 47/33 you can load it a bit and have better performance, lower noise, better psrr, make a better layout, and runs cooler. Not worth the small price difference unless you really need that space. But to each his own, we all prefer different things 🙂
It's not about size (although sometimes that's important too).
Sure, but I used 47/33 when a sufficient current requires, but not for just a few OPs.
An ultra-low noise power supply for an OP is a whim. For few OP (<50-100mA), TPS7A39 is very good, also M5230E (<30mA). For OP power in many applications, even 78xx/79xx or 317/333 are more than enough.
Where extra low noise is really required (for example, reference signals in top ADCs/DACs), a good solution is the LT3042 or any other regulator (including 78xx) + RC filter + OP with sufficient output current (AD8397, LME49720, etc. .).
Alex.
P.S. The big advantage of 47/33 is the fixed output voltage, which is set by pins and not by resistors in the feedback.
But this can also be a disadvantage if you need to change it on an existing board, and the pins are already hard-soldered.
I'm not against 47/33, and I use them in many devices.
Plus, for beginners, soldering msop-8 with a thermal pad (in the case of tps7a49/30) can be easier than QFN.
Last edited:
We are talking here of an off the board reg to replace existing ic solutions. It is heaven and earth difference how much better it is in that use case. Universaly much better reg choice. Even on board, i choose it (it's diy, we can, we are not mass producing tens of thousands devices wherr every cent counts), sounds better, performs better, heats less, costs insignificantly more for our hobby. If we went full EI on this, then we would cut corners, make compromises, but i don't see the point, as again this is diy.
As for dac chips yes, they are more succeptible than ops, but one can argue that lower zout tends to have have better impact than little bit more psrr (i tried both lt3045 and lt3092, comparing them to tps7 on the same dac chips, and i audibly liked tps better, hence why it stays in my dac designs). And many have had this experience as well, but lets call it subjective and not take it into consideration here. But i do like to both have technical aspects and audible as my final judgment.
Soldering wise, no, never use hot air gun to solder components, that is a very very bad practice. It is a tool to DESOLDER and rework components, why it ended up being a choice for soldering beats me.
As for dac chips yes, they are more succeptible than ops, but one can argue that lower zout tends to have have better impact than little bit more psrr (i tried both lt3045 and lt3092, comparing them to tps7 on the same dac chips, and i audibly liked tps better, hence why it stays in my dac designs). And many have had this experience as well, but lets call it subjective and not take it into consideration here. But i do like to both have technical aspects and audible as my final judgment.
Soldering wise, no, never use hot air gun to solder components, that is a very very bad practice. It is a tool to DESOLDER and rework components, why it ended up being a choice for soldering beats me.
What you use to solder just one chip like 47, not all board?
(I have smt oven, but for one chip I many years use hot air gun)
I use soldering iron for single chip (when is that the case anyway?), much much better (not easier, just better), way less of a chance to destroy the chip and surrounding components. Do note i already wrote, for diy purposes i use TH pad below for gnd/thermal, so belly is also solderable.
A Tefal can be nice but you need to check the temp and use a low temp solder paste. It's 101 diy though.
Don't know, why it is better.
I've soldered (at home and at work) hundreds of QFN chips using a hot air gun (not just TPS7A47/33) and never had a problem.
W/o stensil and w/o soldering paste, just put a solder on pads and don't skimp on flux.
(Including bigger QFN like XMOS and other CPU, soic8/msop8 OP, etc.).
Alex.
Last edited: