Development of a "reference" class D starting point

Status
Not open for further replies.
Jump to Latest
#121
Guys, you so fast, that I haven't complete trace of this thread , but here some tips about SMD using:
A) Just try it (and keep trying) 🙂
B) buy the thin tip iron.
Ñ) Try a paste flux (don't forget about finally cleaning the PCB, if this paste are "low active, cleaning isn't necessary" then do it twice🙂, to avoid pin to pin shorts, that even isn't need to soldering SOT23, 1206, SO8 and others our typical class D packages.
D) My buddy used ordinary aquarium compressor+needle of the syringe (like a tiny vacuum cleaner) for taking and mounting SMD, when didn't want buying the soldering station.
 
#122
classd4sure said:
Doesn't hurt to give a new/different program a fair shot every now and then, to be honest if I keep using it a bit more and find it's not very hard to add competitor models to it, the ease of..or perhaps...fluidity? of it's use will have me sold.

I just took a look at a motherboard I have laying around, WOW...they sure make em small don't they? Since you have more knowledge than I, perhaps since this isn't going to be headed for a robotic assembly line, would it be possible to limit ourselves to a certain size (the biggest available) of SMT, if so, how bad would that be to work with? I'm extremely skilled when it comes to working with my hands but I just don't see myself soldering anything the size of a breadcrumb, if I can see it or not.
Click to expand...

I am not sure how the model supply will progress with LTspice either. We can hope that it will be bountiful.

I have done poorly with SMD having smaller lead pitch than that of S0-8. But, I do everything p2p since I never have a final circuit to which I do not continue to make changes. Not being able to keep the circuit two dimensional can be bad concerning noise, though.

IVX said:
Guys, you so fast, that I haven't complete trace of this thread , but here some tips about SMD using:
A) Just try it (and keep trying) 🙂
B) buy the thin tip iron.
Ñ) Try a paste flux (don't forget about finally cleaning the PCB, if this paste are "low active, cleaning isn't necessary" then do it twice🙂, to avoid pin to pin shorts, that even isn't need to soldering SOT23, 1206, SO8 and others our typical class D packages.
D) My buddy used ordinary aquarium compressor+needle of the syringe (like a tiny vacuum cleaner) for taking and mounting SMD, when didn't want buying the soldering station.
Click to expand...

Hi Ivan,
Those tips are useful. They correlate well with the experiences I had when I used to replace SMD microprocessors in consumer stereos and televisions. Alas, in such repairs, the replacement usually did not solve the problem but it was helpful to eliminate the possibility before doing further troubleshooting.
 
#125
If one uses one of those sexy high-speed IC comparators he has to do level translation in order to control a driver IC like a IR2110.

Since a slower comparator seems to work better for an UcD AND one has to use level translation anyway I think it is better to use a discrete one from the beginning.

Regards

Charles
 
#127
My preferred topology would look that way approx:

One dual opamp to 1.) build a differential input and 2.) increase the loop gain by the use of a lag filter (as already mentioned).

A discrete comparator/level translator, which doesn't necessarily have to be the circuit from my simulations.

An driver IC like the IR 2110.

A pair of N-channel mosfets.

Output filter using some air-gapped ferrite core (some RM or "double E" core with insulation material inserted).


Regards

Charles
 
#131
phase_accurate said:
If one uses one of those sexy high-speed IC comparators he has to do level translation in order to control a driver IC like a IR2110.

Since a slower comparator seems to work better for an UcD AND one has to use level translation anyway I think it is better to use a discrete one from the beginning.
Click to expand...
A slower comparator only works "better" in the sense that it keeps the free-running switching frequency down. (There may also be more or less issues to deal with concerning de-coupling and PCB-layout, depending on the high-frequency gain of the comparator.)
Personally, I prefer as high frequency as practically possible, as I favor sound quality over efficiency.
That said, I'm guessing *this* community is likely to want around 500kHz (?), and not 2MHz (!).

Does someone have, or could someone make, a simulator circuit with an integrator circuit (op-amp-based), followed by a discrete comparator, followed by the discrete UcD output driver? (LTSpice or PSpice.)
I'm wondering about what transistors to use for the driver; similarly to subwo1, I am alo concerned about the current delivery capability of the driver circuit. (Maybe that's not a big problem when driving "small" mosfets though?)
I hope to play around with this in the weekend...

As for the SMD debate, I really think we should avoid SMDs for the "start here" circuit, to not shun away all but the experienced DIY:ers. Remember, this was supposed to be a "start here" reference design, not the end-all, be-all.
Already now, I can see it forking into several future variants; one really high-power/high efficiency with a lower switching frequency, and then one aiming for high-end sound with not as high efficiency and power capability...

Regards / Johan
 
#132
John you are completly right. The delaytime of the complete amplifier makes the whol thing goes in oscillating. Lower delaytime increases frequency.

I would go for a fast discrete comparator, those small fast switching transistors like BC546 will already do the job.

What is the problem with SMD?
I use a Weller WS81 solder with a normal tip, normaly we use only 0805 SMD --> no problem at all.
To solder the smd, I put a little bit solder on one of the pads of the pcb, pick with a small tweezer the smd 0805 resistor, put with the tweezer the resistor on top of place of the pads. One of the pads has a little bit solder, heat this solder very quick. Voila the smd resistor is connected with one solderjoint. Now solder the other pad and eveything is ready!

This goes even easier as with thrue hole components.


Regards,

Jan-Peter
 
#133
Does someone have,
Click to expand...

no



or could someone make,
Click to expand...

yeh

I plan on adding the differential input before I work on the extra integration loop though. I'll see what I can get done.

I don't think the current capability of the driver stage is a problem at all. I already made one capable of three amps and it simulated switching a pig of a mosfet at 1 amp just fine. It was designed to handle 1 amp, used half that, but that can and will be improved. I'm not set on which transistors yet, it will depend on the mosfets I pick. Pafi was so kind as to recommend some very nice ones which I haven't yet implemented but I think I will. I'll work on it.

I'll do it with pspice though.

I'll go with ditching the SMT for now as well, we'll keep the dual layer board though, call it a comprimise. If anyone pulls a trace off they can make themselves another board easily enough anyway right.

I dont' see a problem getting that done but I might require some advice along the way. I'll keep you bothered.

In the end I'll be going for extreme power and fidelity, but I'm not convinced you need to switch faster to get extra fidelity. 500khz sounds good to me for now anyway.
 
#134
Personally, I prefer as high frequency as practically possible, as I favor sound quality over efficiency.
Click to expand...

Why not try to have both ?? I am convinced that high-quality amplification is feasible with lower switching-frequencies than 500 kHz.

Since my demo version of P-SPICE is very limited I can't help you much with your question, otherwise I'd have done it.

I don't have models for decent & modern OP-AMPs so I usually use the ideal voltage source for building integrators, active filters and such.
Comparators can be built using the ETABLE element.

Regards

Charles
 
#135
What is the problem with SMD?
Click to expand...

Nothing, it's just that alot of people might find it very intimidating, we can make it clear to them that through hole isn't the way to go for this and it's done for their benefit. Project 2 will be all out 🙂

STickin to the plan..

I still like that HIP2101 too, but there's more to learn from a discrete version, and easier to simulate if the design is given. Keep this circuit as a launch pad for anyone interested.
 
#138
Ok I have a part or two that might make things noticeably simpler but first I need to confirm the gain of the input differential amp stage? From my simulations and the fact Veb cant' be much more than 5 V am I correct in saying the input to the discrete comparator cant' be more than 5V pk to pk? If so would a gain of 10 for the input stage be too crazy?
 
#140
Hi,

Actually I meant to say 2.5V peak, as Veb can't be more than 5V. Anyway I'm refering to the actual audio input signal, not yet mixed with feedback or anything else. What should the gain of the instrumentation amp ideally be? I have a 1 part Burr brown instrumentation amp with built in laser cut resistors matched to 0.01%. Nice huh...gain is fixed at 10. I understand Jan peter uses a gain fo 5 at this stage....it would only require a lower input signal, or am I wrong here.

Regards,

Chris
 
Status
Not open for further replies.
Top Bottom