I absolutely have the skill to do rework and I can buy the equipments if I want to. My advice is from my own experience. You have to admit that working on SMD is 10 times as hard. Of cause it's doable, just slow and have to be very careful. One twitch of your hands and you create solder bridge and again it's lot harder than TH to fix.
I have no problem spending hundreds for equipments, in fact, I have a Metcal station ( should say had as it burnt. It uses RF heater that is not the most reliable). BUT from reading posts, there are people that cannot afford to spend that kind of money. You don't want to make it so rosey that people design with SMD just to find out they have to invest a lot more and screw up a few times.
That's all kind of missing the point. I talked in the last few posts that you can easily find TH parts that are very good, don't get stuck on "the" part. It's all about the design. I would go the SMD as the very very very last resort.
Last edited:
6ppm @ ?
BTW: I hate non TH. I'm getting older and can't see and hold the small devices ...
BR, Toni
P.S.: Hardly waiting for the 2nd Edition ...
It's actually easier depending on what you're doing. If you are home etching it eliminates a lot of drilling. The self aligning of parts in the toaster oven is nice too. I just need to figure out how not to brown the silk screen on my fab house boards now. It's all in what you are accustomed to doing.The most tedious part I find is pulling all the passives out of the reels.
Not necessarily a problem, and usually a benefit. At minimum, it increases the class A region, when optimally biased, by a factor of 4 in power.
Not sure if you were referring to BJTs or MOSFETs. The latter need to be matched to some extent, which is usually not difficult for parts that came from the same tube.
Cheers,
Bob
6ppm @ ?
BTW: I hate non TH. I'm getting older and can't see and hold the small devices ...
BR, Toni
P.S.: Hardly waiting for the 2nd Edition ...
Hi Toni,
Yes, that is 6ppm. I believe that was THD+N with a 200kHz residual bandwidth.
Cheers,
Bob
Hi Pete, All,
One thing about surface mount that will not change is something we touched on. The heat, smt parts are not rated as high as their TH prototypes.
So, who wants to be forced into using Al core PCB material? Then to my mind we have a possible contamination issue. Pin spacing is far closer so that leakage currents might also be an issue in certain situations. How many of you don't clean off your flux residue?
-Chris
One thing about surface mount that will not change is something we touched on. The heat, smt parts are not rated as high as their TH prototypes.
So, who wants to be forced into using Al core PCB material? Then to my mind we have a possible contamination issue. Pin spacing is far closer so that leakage currents might also be an issue in certain situations. How many of you don't clean off your flux residue?
-Chris
I must admit that I have looked at SYN08's (Ovidiu Popa) simulation of EF stability but don't completely follow it.
I haven't noticed you reply to Richard, so I would add my plea to his.
Just about to finish a contract so I want to return to amplifier stability.
If you don't want to hijack this thread then post in one of mine, or your own.
Best wishes
David
On this subject - I did it again ...
Compare the LT output -
http://www.diyaudio.com/forums/solid-state/248105-slewmaster-cfa-vs-vfa-rumble-159.html#post4369547
To the real thing -
http://www.diyaudio.com/forums/solid-state/248105-slewmaster-cfa-vs-vfa-rumble-160.html#post4371603
In between those posts I predict with just the simulator how good it
(might be). The real output actually was better than the simulated one
at 100K. The builder lacks FFT or an actual AP , but I bet that would
also not be too far off.
Does 100K and overload WAY better than a blameless - (horrible ,
evil symmetrical design)
OS
Yes, it looks like it clips very nicely.
But you wrote -
And Richard asked for an actual ASC example to see what you actually did.
I would like to see this too because that would be the clearest way to explain and for me to learn.
Best wishes and keen anticipation
David
Hello everyone,
SMT is a skill, it requires lots of practice. If you are not going to invest the time or $ then you are forced to get someone else to do it for you. Gmarsh's "Weiner", I'd say he soldered up the majority of them for the DIY group buy, so that shows where the smt skill levels are at in this community = (scared/kindergarden), so I will try to help to improve them in the DIY forum if folks are willing to give it a try.
For bad eyes, you have to invest in good magnification. For a steady hand, well you have learn to rest your hand against a solid object, for stability, your fingers do not shake near as much. My Dentist has a real cool set of second eyes
I have a cheap mag head band, big eye mag with light, 10x jewelers loup, two Weller irons, solder wik/sucker, glow core solder, tweezers. Having lots of light is so important too.13368 AIM | Mouser
Member "OPC" highlighted the Susumu thin film resistors to me, so I use them
These seem to be a good deal
RR1220P-241-D Susumu | Mouser
I do agree use 0805 or even 1206/1210 if you are not pressed for space.
I am working on building the wire (LME49830 Latfet) amp board of late. It will use the PS from an old Pioneer SX-1050 receiver (120W/ch) mounted in an old Sansui G-7500 chassis. Also plan to put in my media player/ Si4735 or Si4770 tuner and some remote control for power volume and probably channel selector. OS and jwilhelm are working on the Pitchfork pre-amp so that is a possibility too, to add the SPDIF rx and DAC. Another retirement project
Breadboarding is great/fine, this was our DIY staple for years since a pcb was so expensive. Now with low cost Asian pcb fab, it can be more cost effective (time/$) to do a pcb but that also means you need to know pcb layout too.
Cheers
Rick
Last edited:
No I do not agree with 10x as hard. It requires different tools. I think it the other way around on assy time. Try this scenerio, a 1/4W R, THT you have to lead form/bend, insert solder,clip leads. SMT, dab of solder one pad, place with tweezers, solder other lead = less time, on my stop watch.
It really depends on the type of design, for sure but for the majority of my designs, they would not be possibly since many comps are only available in smt. smt saves expensive pcb area too.
This particular thread on power amps, well talking smt is rather a bit off topic because it is not necessary to do in smt as we are into high power devices and devices are available in tht. One would only consider smt to save space for some small signal/lower power such as IPS/control/protection and THT exclusive or THT/smt mix for OPS.
This is your opinion, and I disagree, I want to let people know the other side that working on SMD is not that easy, from a real engineer that is in the industry. I work a lot with SMD for years, I know. I encourage people to desolder a SOIC8 and resolder on using a dead board and let them be the judge.
You might be exceptional. I can say I am very good with my hand eye coordinations, I can do it, I have breadboard with 0603 and IC for 2.5GHz SONET receiver circuits on a copper board. I know how to work with SMD. Not all people are as good as you. Let them try first before investing in expensive equipments. I would not advice people to use toaster oven and all that, it might be dangerous. In an open forum, you do have to be careful in advising people using unconventional way when there is a chance of injury. I can only advice using proven tools, that is heat gun with special tips, dual solder station, very good magnifying lamps. Don't treat everyone as an expert on the forum.
Of cause, I am talking about IPS and VAS with small signal transistor. And yes, read Cordell's analysis and do the design and compromise. I am willing to bet people can't tell the difference between circuit that people so picky with "the" transistor vs one with good through hole transistor and GOOD design.
Last edited:
Reworking an SOIC8 is easy, even with just a soldering iron (remove as much solder as possible with solder wick, then thread a very thin wire between the package and the legs, then heat one pin at a time and pull the wire through underneath the pin being heated, to lift the pin from the board.). I can rework MSOP8 (with hot air for removal stage) without needing magnification. Clearly, it depends on the person, but I think you are making SMD sound a lot harder than it really is. With a hot air pencil, Reworking SMD is often easier than reworking TH in a PTH board where getting solder out from inside the plated hole can be very difficult and sometimes impossible if the component lead is a snug fit in the hole. With a TH transistor sometimes you need the same tools for rework as SMD, namely a hot air pencil so you can heat up all 3 legs at once and pull the transistor out.
Suggest this method as well, it is faster than the hot air machine, that is, once you make the simple apparatus, but that took me a few minutes. i made up two of them so far, once on a SSOP(Si4735-D60) and then on a SOIC.
http://www.diyaudio.com/forums/group-buys/269855-wiener-tpa3118-amplifier-card-23.html#post4258571
http://www.diyaudio.com/forums/group-buys/269855-wiener-tpa3118-amplifier-card-23.html#post4258571
I'd go along with that too. I've worked on boards with standard DIP package devices that are so tightly fitted into their plated thru vias that removal is extremely difficult. Even snipping the leads and extracting one at a time isn't always as easy as you might think because the via is so small you just cant easily clear the solder. Stainless wire can be useful for probing a way through in such cases as the solder will not take to the stainless.
Some of you will be familiar with this,
http://www.diyaudio.com/forums/parts/127924-working-smd-how-do-without-specialised-tools.html
which shows what you can achieve with a poker and braid
Must be a lot of them, based on what the distributors are stocking. For small simple single sided pcb's, el cheapo stuff like your smoke alarm, co detector. Take one apart and take a look.
More about smt processing on the cheap
Seeed Recipe - Hack a Toaster Oven for Reflow Soldering
Ha ha, if I producing audio product, I definitely use SMD as much as possible. smaller pcb($$$), likely cheaper for production as it's pretty much computerized and done by machines. And not for anything else, prevent people with itchy hands from getting into the circuit. It's easy to mess up and they can't return to you for warranty repair!!!
May be not for audio frequency, SMD is so much superior in performance in every which way. You can eliminate a lot of the parasitic and EM emissions.
But even if I do prototyping for audio, I still prefer TH. It never even cross my mind not to put socket for DIP. I was even thinking about putting pin sockets for TO92 transistors!!! I did put pin sockets for transistors in the mid 80s at work for first run on ultrasound medical imaging that was 20MHz BW with 1GHz GBWP. It make changing transistor that much easier. those pin socket fits in 035 holes and it fit almost flush to the board. I decided against it as I think BC5xx and KSA992 and KSC1845 will cover all my needs already and socketing a 3 pin device is not reliable and can shake loose.
May be not for audio frequency, SMD is so much superior in performance in every which way. You can eliminate a lot of the parasitic and EM emissions.
But even if I do prototyping for audio, I still prefer TH. It never even cross my mind not to put socket for DIP. I was even thinking about putting pin sockets for TO92 transistors!!! I did put pin sockets for transistors in the mid 80s at work for first run on ultrasound medical imaging that was 20MHz BW with 1GHz GBWP. It make changing transistor that much easier. those pin socket fits in 035 holes and it fit almost flush to the board. I decided against it as I think BC5xx and KSA992 and KSC1845 will cover all my needs already and socketing a 3 pin device is not reliable and can shake loose.
Last edited: