From: info@turneraudio.com.au   
      
   On Tuesday, 11 February 2014 11:04:22 UTC+11, hugeshows  wrote:   
   > Hi All,   
   >    
   >    
   >    
   >    
   >    
   > This question pertains to the Sherwood S-5000 restoration I'm doing.  I am   
   interested in getting to the bottom of a power supply resonance I believe may   
   lie in the stock S-5000 power supply.  I've been noticing lately, and have   
   noticed in the past the    
   tendency for this amp to make the speakers flub a bit at an extremely low   
   (1-2hz) frequency when there is a transient with a large attack, and resulting   
   large instantaneous current draw.  I had first thought that it was present   
   only when the amp was    
   driven near its clipping point, but have since noticed it even at normal   
   levels, though it is much harder to notice then, requiring you to put your   
   finger on the cone.  I even recall in my dark memory of other S-5000s I've had   
   that one amp was even self-   
   oscillating at idle at this same approximate frequency when the original power   
   supply caps got really weak.  So it appears to be a bit of a design flaw in   
   general with at least the second version S-5000.  I'm not sure if I've seen   
   the first one do this    
   or not, but I am going to find that out shortly anyways!   
   >    
   > First of all, I am still in need of a new scope since my 535a found a new   
   home, and I am unable at the moment to snoop beyond the resolution of a VTVM.    
   So I have to resort to a more primitive mode of diagnosis at the moment.   
      
   For goodness sake, spend some money and gear yourself up with a few more volt   
   meters able to measure 0.1Hz to 200kHz and a dual trace CRO!!   
      
   How the hell can you observe electronic behaviour without tools?   
      
   And how the hell can we understand your problems and your questions if you do   
   not refer us to a reliable working website address where a full legible   
   schematic exists of the gear you cannot understand?   
      
   However, be those shortcomings as they are, I see a few possibilities for the   
   cause of LF instability in your amp....    
   >    
   > However, there are several reasons why I think it may be a power supply   
   resonance, and please feel free to shoot me down if you think I'm on the wrong   
   track.  First, the oscillation occurs in both channels with the same intensity   
   and always in phase.    
      
   Both channels would have identical schematics and NFB networks. There is a   
   chance that the stupid designer did not check the F response between 0.1Hz and   
   200kHz after completing a prototype and there may be a big peak in LF response   
   due to the C&R    
   couplings between stages and the OPT inductance and load R.   
   At very low F below 5Hz the NFB can become positive FB due to phase shift.   
   Maybe your amp needs an additional R&C network placed between V1 and V2 amp   
   stages. Usually V1 has a 0.47uF feeding V2 grid with say 220k grid biasing R.   
   This looks OK but you need to reduce the open loop gain ( gain of amp without   
   any NFB connected ) below 20Hz. So I suggest you get a 1M + 0.05uF and connect   
   them in parallel, then insert them between output of 0.47uF and top of 220k   
   which is V2 grid. The    
   OLG will then become about -12 lower at below 5Hz and the LF reponse will lose   
   its peak if there is one.   
      
   Its also possible that the original failed miserably to put large enough cap   
   values in the B+ rail and such mal-practice led to many old amps motorboating   
   just slightly because they are on the verge of full oscillation.    
      
   Many old amps would oscillate at LF especially when no load is connected. This   
   is because the gain of output tubes is highest with no load and the amount of   
   NFB applied is increased when open loop gain is increased. So an amp with a   
   speaker connected has    
   the speaker winding resistance across the output which usually provides a low   
   enough load to prevent OLG rising just enough to allow oscillations. But you   
   DO HAVE a badly designed amp, and I bet there are many other things that   
   should be dealt with, not    
   just what you are finding.   
      
      
    I have yet to see this occur in one channel and not the other.  If I set my   
   balance all the way to the right, I still see the oscillation in the (silent)   
   left side.  So all of that tells me that plate voltage is oscillating when   
   transients hit,    
   regardless of channel overload condition.   
   >    
   >    
   >    
   > Since tube rectifiers have a much higher impedance than silicon diodes, I'm   
   wondering what can be done to eliminate or greatly damp this resonance without   
   throwing in the towel and re-designing the whole power supply for solid   
   state.  It's not like    
   there's a ton of room in this amp to add a choke and a bunch of big caps,   
   though I'm sure I could fit in a few well chosen components if need be.  It's   
   pretty clear that this amplifier is capable of doing even more than it is now,   
   and it sounds pretty    
   damn good as it is.  But having the HV+ sag enough to actually couple through   
   the output iron at ~1hz is both a testament to this iron and an indictment of   
   tube power supplies, at least this one.   
      
   At least you should be able DUMP the tube rectifier. They are bad things in   
   compact amps because they cannot stand high peak currents, and just fuck about   
   getting hot and making things around them hot.   
   But if you replace say a 5AR4 with a pair of IN5408, then the working B+ bay   
   become 15% higher than wanted, so you need to place a pair of series R between   
   each end of HT winding and Si diodes to mimic the plate resistance of the tube   
   diodes. These R    
   will get hot, but the filament heating of the tube rectifier will be missing.   
   So maybe 150r x 10W will do, but they must be tried and measured until   
   correct. THEN, the reservoir cap can be a much larger value than whatever it   
   was with tube rectifier.    
   Surely there is room for such cap replacement.    
       
   > So, here's the circuit in a nutshell, stock and as it is now.  I should   
   mention that while a couple cap sections have increased in value now, the   
   oscillation hasn't really decreased at all.  No surprise, the resistors and   
   limited value for the first C    
   are probably to blame.   
   > Stock:   
   > From the 5AR4, first 20uf, then 33 ohms, 40uf, 4.7K, 40uf, 10K, 30uf, 56K,   
   20uf   
   > The plate supply comes from the junction of the 33ohm resistor and the 40uf   
   section.   
   >    
   If the wanted working B+ at the reservoir cap afer diodes is under +400Vdc,   
   maybe you can use 470uF x 450V rated, so that during turn on the B+ does not   
   soar much above +470Vdc while tube heaters turn on.    
      
   [continued in next message]   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)