Hi Number_47,

I didn't see any replies so I will give a try.  Going to local live shows and see how other guitarists get their sound, along with getting their opinions, is a good way to decide what to buy.  A side benefit is that it helps them have an audience.

I have seen a lot of live shows in the past 3 years at local venues and it is rare to see an SS amp (other than the bass player's amps).  Usually the guitarist carries in a tube amp of some kind, so I guess the tube reliability & maintenance issues are acceptable to most players.   The worst SS amp I saw used by a working musician (just a local beginner) was a Line6 Spider - it was overloaded and very buzzy.  The most recent (that I noticed) was an old SG System, but the player was just using it for "clean" and his sound wasn't very prominent in the mix so I couldn't judge the amp well.  There have probably been some Peaveys and some Marshall SS in some shows but I don't know the models well so I can't say for sure if they were SS.  Probably only 5% of the guitarists in local bands & regional bands use SS at the place I go, where the music is usually alternative / classic rock / rock-rap / lite country.  I see mostly tube Fenders and Mesas, with perhaps as much investment in pedals as the amp.

The discussion, "Solid-state guitar amplifier history -thread" http://www.ssguitar.com/index.php?topic=469.0 has a good list of various SS amps, including old ones.

Try attending live shows of other bands, and when you  see and hear something that is attractive, ask questions of the guitarist after the show.  If you are settled on SS and serious about performance I would recommend buying a good quality new or used SS amp rather than trying to build.  You don't say enough about how you use the amp (effects pedals, music style, how many guitarists in the band, rhythm versus lead, reverb, whether it is miked to the house PA) to answer better.

Once I thought I saw a "home built" amp in a performance but found it was really some small company's effort - it just looked very crude.  I forgot the amp's name - it was tube amp.

Hi phatt,

As far as what I am chasing, probably it is what happens with most people who mess with guitar amps.  You read reviews, get a cheap example, it is OK but not perfect, read more, get another amp, it is OK but not great, read more, and so on.  Several months ago I was trying to build circuits and see how they sounded but can't get motivated lately.  The amp mfgrs already have so many excellent products, and right now prices are very good, so I am sure I won't be able to rival them.  But it is fun to learn about the concepts... maybe I'll get motivated again.

Thanks for sharing your views.

Hi Teemu,

One thing I wonder about is whether the marketing of "30W versus 60W" might someday be less important.  There seem to be a lot of sales right now of the little 5W tube amps like the Valve Jr and Blackheart.   Even with the higher cost amps, the Mesa Express 5:25 / 5:50 and Egnator Rebel are selling because people realize it just doesn't make sense for most people to own a 40W or 50W tube amp.  Maybe something like this will happen for SS.

Hi phatt:

You asked, "why not use an output transformer?"  Good question...

I am not sure if it would help or hurt.  It seems like it would be very good at preventing DC offsets from getting to the speaker and damaging it.  Instead, the amp's transformer would become damaged.  At least you wouldn't be frying speaker coils due to a transistor getting shorted to the power rail, so that's some progress.

I know that old amps like Acoustic used transformers, but maybe it was because they were limited on the availability of transistors (only NPN? was available for awhile).

The hi-fi amp books I have read argue against output transformers.  I have Doug Self's book and Slone's book.  But their interest is sound reproduction.  The cons are: transformers add cost, aren't really needed, cause phase shift, make the amp more narrow band, add weight and mechanical complexity.  Certainly you never see output transformers in newer SS amps built by Fender and Tech21, etc.

I have seen the theory that transformers have some ability to saturate non-linearly, and that is part of vacuum tube tone.  I think there is even some patent where the author introduces DC offset on purpose to try to imitate tube tone.  I don't know if these things are provable or just wishful thinking.  I know the some of the tube people think that it's helpful to use a high class transformer to get the best sound, and there are differences in the construction that are important.  Mercury Magnetics sells an upgrade transformer kit for the Valve Junior for asking price of $299... seems like a lot of money for an amp that can be seen on craigslist for $100.

Do you have a theory about output transformers?

I noticed a couple of more questions in replies to my original question:

Hi Brymus:

You suggest "why not ditch the V-I limiters and just use a thermal shutdown device?  Overbuild it".

I think this is the general way I will proceed.  I need to study more about the causes for the failures, though, since the failure mechanisms might be more than just simple overheating.

Hi gbono:

You ask "Is the issue that the input signal into the amp should not have any "non-linear" processing so one can overdrive the OP section? or is DSP still to costly? ??"

I like the idea as you state above that avoiding preamp clipping would allow you to have a more flexible input section.  I think it would be nice if the preamp could handle any reasonable amount of voltage that you might find in a pedal. 

As far as DSP, I want to avoid overcomplication.  DSPs require a lot of tools and different skills than working with analog circuits and the DSP chips themselves are "here today, gone next year".  Right now the DSPs and high power processors tend to be going toward very low voltages and even require a couple of different power supplies.  And then you are working with tiny chips that require a professionally made PCB to be reliable (think of BGA packages).  So pretty soon you are not even working on an amp - you become distracted by the digital stuff. 

And anyway, there are no shortages of DSP based designs like the Line6 that have already taken things pretty far.  I have the Line6 XT Live Modeler and it is pretty good for what it does so I don't have an interest in trying to duplicate it.  It works pretty well but it does have limitations.

Hi Teemu,

The articles I have seen about the lateral MOSFETs seem to say that the parts are not very popular any more, and have become rare.  My ideal design would include very basic parts that anyone can find (from a place like Digikey or Mouser or Newark, etc.), and that you would still be able to build it 15 years from today.  No cadmium photocells, CA3080 chips, germanium, rare transistors, etc.

Hi R.G.,

Your post is very informative.  I agree with the comment that the V-I limiter is an economy move.  Your suggestion of having everything over-built to the point that you can handle shorts / opens / transient shorts / odd impedance seems like a good approach to me.

Certainly we know from reviews of tube amps that they also have problems.  I remember reading that some of the tube amps that include circuit boards can develop permanent damage where the arcing (due to having too high load impedance or open circuit) causes the tube's PCB to get carbon burns that cannot be repaired.

I want to try to design an SS amp that has some reasonable limiting on the preamp stages, but it is very hard to "softly" clip due to the crest factor problem.  The amplitude of the guitar signal's peaks are very large compared to the average signal.  So I haven't figured out any good approach that seems worth building.  I have tried the easy circuits (like diode limiters) that you see in the old schematics but they always sound very bad compared to a good tube amp or the digital processed modelers.  Anyway, I keep looking.  Thanks.

Hi Teemu,

Thanks for your links to the patents.  It seems like the goal in them is to sense temperature and back off the drive till danger has passed.  That seems like it might work.  It seems very complicated, though.

...It doesn't much matter what a damage-prevention method sounds like, if not having it means smoke and flames  - or just dead, permanent silence - pouring out of the amplifier box...

...would you rather listen to the V-I limiters sometimes inserting a bit of grak or to have to send the amplifier into the shop to have the output transistors replaced whenever the V-I limiters operated?...

Hi R.G.,

I am not sure the picture is as back/white as you paint it.  I was hoping someone would reply with a link to some superior limiter design, something that was better than the typical V-I limiter.  I think we all recognize that vacuum tubes don't use V-I limiters, and when overdriven into clipping they do not usually cause smoke & flames.  So it seems to me there ought to be a SS limiter that avoids "grak" yet also protects against damage to the outputs.

Does anyone have links to a better design than the usual V-I limiter?  I think the V-I limiter would degrade the sound of a guitar very much.  First, even in an ideal simulation you can see it cuts off the tops of the signal, making distortion that is obvious.  In the real world the V-I limiter would also cause DC imbalance because the components that limit the positive peaks are going to be different than the negative simply because of part-to-part tolerance variation.  Also in the real world the music signals are not symmetrical sine waves that would clip evenly.  Instead the positive peaks of the music signal might be very narrow while the bottoms are wider.  So you might only limit the positive peaks, thus causing another source of DC imbalance.  The DC imbalance is not good for the speaker, so the supply will be affected, and this will lead to further degradation of the limiting. 

I also don't like the idea of the V-I limiter because you have two control loops fighting for control of the signal (the global negative feedback versus the local control of the limiter).  It seems like timing problems will arise during transitions (into & out of limiting).

So is there a better method?