This is a thread from Synth-DIY on matched transistors and rbe compenation. I'm in the midst of ordering a few parts so i can finish up my ASM-1 (finally). I've decided to go with the MAT-02 and tempco option. Any idea on current catalog numbers or other names for these... I did a search on their site and couldn't find anything for the MAT-02.... lots of npn transistors though. Help please Nicholas Pelkey ---------------------------------------------------------------------------- If you go to Analog Devices website and ask nicely they will send you 2 MAT-02's for free! (samples) And they arrive quickly too. You do have to sign up with them & give them an email address and some other details...nothing too serious though. Peter Snow ---------------------------------------------------------------------------- Yes, but are these really neccesary, or can two standard matched trannies epoxied together do the trick? "Rob" ---------------------------------------------------------------------------- Seems like I remember some list member measured the Bulk resistance of various off the shelf transistors. The one nice thing about the MAT02 is the low bulk resistance. If somebody is going to match a pair of regular transistors, I would think that picking a type with low bulk resistance would be the best way to go.... patchell ---------------------------------------------------------------------------- You can use standard transistors, find a matched pair, glue them together and use that. Sure. It's all a matter of deciding if one wants to do the whole procedure or just get a ready-made. The closer matching you want to do, the more troubles you have to go through. One has to have a good measurement setup (either in a DMM or people roll their own, like the well known Moog matching setup), one has to have a relatively many transistors in order to find a good matching pair and one also needs to be carefull when handling the transistors (they need to be at the same temperature, so fingers may heat them up diffrently). I personally prefer to spend the few hours that I have for DIY doing other things, but I have not ruled out matching either. I've even spent money in order to be well prepared. Magnus Danielson ---------------------------------------------------------------------------- > A couple of us had a look at that a year or so ago. > > Thing is ... you can compensate the small bulk-resistance error exactly with > simple standard techniques, so it doesn't really seem to me that it's worth > the expense to buy super-low resistance units. Thing is, most expo-converters does not compensate for the bulk resistance. I haven't dug into the details of bulk resistance compensation myself, but maybe somebody och something will kick my lazy but to go and do the reading and theory exercises. Now, is there a simple mod on a standard converter? Also, this brings up the topic, how does one matches transistors for a expo converter as well as how does one best trim the bulk resistance compensation? If we where to use the ASM-1 expo converter (which also exists as a separate schematic over at my schematic page) and discuss that one... Magnus Danielson ---------------------------------------------------------------------------- > > Thing is, most expo-converters does not compensate for the bulk resistance. > > I haven't dug into the details of bulk resistance compensation myself, > but maybe somebody och something will kick my lazy but to go and do > the reading and theory exercises. > > Now, is there a simple mod on a standard converter? > It just takes a diode and a trimpot. See any of the VCO's at my site. For example the second figure at http://home.earthlink.net/~ijfritz/sy_cir4.htm "Ian Fritz" ---------------------------------------------------------------------------- But isn't the text incorrect? It says "the the D1-R14 feedback path" Drop one "the" and change R14 into R13. The downside of that solution is the need of an diode which needs to sit at the same temperature as the expo-core transistors. In your solution it doesn't really care, since you use a CA3083. I see a diffrent solution in my ol' PMI databook (hm, what did I just say about those?) where they basically only adds a resistor to the base but to ground. Anyway, I'll have a look at both. Magnus Danielson ---------------------------------------------------------------------------- > But isn't the text incorrect? > > It says "the the D1-R14 feedback path" > > Drop one "the" and change R14 into R13. Ooops! You are correct, of course. > The downside of that solution is the need of an diode which needs to > sit at the same temperature as the expo-core transistors. In your > solution it doesn't really care, since you use a CA3083. Since the correction is small, matching and thermal sinking of the extra diode should not be crucial. But I haven't checked this in any detail. "Ian Fritz" ---------------------------------------------------------------------------- > Seems like I remember some list member measured the Bulk resistance of >various off the shelf transistors. I did that. The results are at http://www.uni-bonn.de/~uzs159/rbe.html My personal advice in a nutshell: Use cheaper pairs and the aforementioned Rbe compensation. To get the optimum HF performance you'll need this anyway. These cheaper duals (like 2SC1583 or the like) will be much better matched than spec'd. Leave those MAT-02s for the audiopath. Glued transistors work, although thermally they're inferior. René Schmitz ---------------------------------------------------------------------------- > > The downside of that solution is the need of an diode which needs to > sit at the same temperature as the expo-core transistors. In your > solution it doesn't really care, since you use a CA3083. > The way I solve the "diode" problem is to use a diff amp across the resistor that feeds the emmiters of the exp transistors (schematic is on my web page in several places). patchell ----------------------------------------------------------------------------