Determining Filter Cap Values

[Ed_Chambley]

In the designing stages of a power section, how is it one determines the main filter capacitance?  For instance, the Marshall Major uses 375uf and 200uf, which is the largest I have seen in a tube amp.  Lots of other other equipment, like computer boards, I have seen much more.  I know how to calculate ripple reduction and know if you change the first filter cap it seems it will tighten up the bass.  Also knowing the turns in the PT has impact on it as well.

I am in the middle of a build.  I am building 2 amps very similar and based on the Dr Z Route 66.  Push Pull, KT88 and KT120.  It is a UL design and I am going to attempt to get close to 100 watts with one of them which is why one will have the KT120.

Looking at the Sunn 200s, which is similar in design, they used 30uf, 20,20,20 with a GZ34 rectifier.

I plan on ss rectification and I want a stiff power section, but I don't know the direct effect of filter capacitance and why certain values are chosen.  I have experimented with different values previously, but I was wondering if the capacitance is too large what problems may occur.

[PRR]

That Sunn is C-L-C filtering.

Say you have dirty water and need to filter it. You can use one big filter or a 3-stage filter. The one filter will either leak fine dirt, or clog-up fast, or else be very large. The 3-stage filter may be a better idea. Or not.

There is an added problem with weak (vacuum) rectifiers and low-resistance PT windings. The peak current into a large cap may burn the rectifier. This is really more a resistance problem than cap-size, but a smaller first-cap is usually the only convenient way to handle it.

The right-size depends on several things. Plate voltage is not very sensitive to ripple until the amp is distorted (which is routine on guitar). G2 voltage is more sensitive, and in UL the G2 gets plate power not a second-drop filtered supply. In push-pull ideally the B+ ripple cancels, for small signals, in as far as tube plate resistances match. Also you can kill a lot of idle-buzz with hefty overall NFB. (Early Champs, no NFB and 2-stage filter; late Champs 1 filter and some NFB.) Also: your speaker system. Ampeg VT40 on PCB had strong hum, but with the open-back cabinet you could not hear the 60Hz at 2 feet (120Hz is harder to kill).

Further: filtering improves at low current. A class B stage can be tuned for very low current (thus low ripple) at idle. Some big-dogs have large buzz at LOUD output (who can hear?), but are quiet at idle.

> UL design ....close to 100 watts

Find a schematic for Dynaco Mark VI. 120W, UL, heavy NFB, living-room clean. Ah, here's a snip (below):

If you run good NFB (gain really goes down when NFB is connected), this will be excellent. If not, it still may be fine in large (noisy) venue or with bass-shy cabinet.

Some Single-filter dart-tosses:

1uFd per mA (so 400uFd for a 4-tube 500V amp) (this may be too generous in hi-volt work)

R*C ~~=0.1 (0.05-0.2)
C in Farads and R is the equivalent V/I of the load. In AB tube amps, roughly half the OT's nominal p-p impedance. So 5F6a Bassman, 4K OT, 2K load, 40uFd filter: 2K*40^-6= 0.04; Ampeg VT40, 6k6 OT, 3k3 load, 37uFd filter: 0.12; transistor amp 50 ohms 2,000uFd= 0.2

[Ed_Chambley]

Thanks PRR, got the Dyanco schematic and part list.  Answers what I needed.  I own a Mark III, but it seems to be 70 watts or so.  I have never checked, but if I recall correctly the documentation states either 60 or 70 watts. 
On the Mark IV, they have a full wave SS rectifier, series of 100mfd/300v to a choke .85H, 400ma, and 13.5 ohm and then to series 50mfd/300.

This is also C-L-C Filtering, is it not?

I am not going to be adding NFB and I would like to eliminate the choke.  If I eliminate the choke, would it be wise to add additional filtering of 50mfd or would you think a dropping resistor alone would be sufficient?

One of the OT I am using is 4300 ohms CT primary with screen taps, 4,8 and 16 ohm secondary. 20 Hz to 20 KHz response at 60W, within 1 db, 30 Hz to 15 KHz at 120 watts. Max DC per side (suggested) 120 ma.  It is actually a replacement Dynaco OT.

Can I use the 4 ohm tap for a low impedance balanced line out with a send pot?


BTW, thanks for the information and the calculations.

[silverfox]

I'm new at this so I could be mistaken but- I wouldn't eliminate the choke. I believe that would effect the clarity of the higher frequency response by allowing lower frequencies to drain the power too deeply. Another possibility would be introducing ripple into the supply on large surges- low frequencies. My best estimate.

Silverfox.

[Ed_Chambley]

I'm new at this so I could be mistaken but- I wouldn't eliminate the choke. I believe that would effect the clarity of the higher frequency response by allowing lower frequencies to drain the power too deeply. Another possibility would be introducing ripple into the supply on large surges- low frequencies. My best estimate.

Silverfox.

Thanks, this is why I ask.  I will look into your suggestions.

[PRR]

At this power level, it is either a gallon of caps or a cap-choke-cap filter.

I'm not a fan of chokes in smaller amps, but up here it makes good sense.

Especially with no NFB. The VT40's 120W brother is straight cap, no NFB, and the VT40's buzz is tolerable mostly because of poor bass in the speaker.