Connecting an extension cab

[JohnH]

Heres a query about connecting an extension cab to a combo.

I have a 40W tube powered amp (Marshall DSL401), which has a single 16 Ohm speaker and two output jacks wired in parallel (the internal speaker plugs into one of them). It has a switch that allows the output to be matched either to 8 Ohm or 16 Ohm. The idea is that you can plug in an extra 16 Ohm cab, on its own with output set to 16 Ohm, or together with the internal speaker, with output set to 8 Ohm. Or one can plug in an 8 Ohm cab, without using the internal speaker.

I have an 8 Ohm closed-back 4x12 cab, containing four 35W Celestion G12L’s. It sounds great plugged into the combo, but very different to the single 12 in the combo, which has a brighter punchier sound which I also like.

I’d like to be able to run the cab in addition to the combo speaker – not because I really need to, but just because it would be more awesome. So here are my ideas on how to do it, for which I'd be grateful for any comments. I don’t wish to mod the amp, nor do anything non-reversible.

1. I just plug the 8 Ohm cab in, in addition to the 16 Ohm internal speaker. The amp would be set to 8 Ohms but it would see 5.3 Ohms, based on the nominal resistances. This is not recommended in the handbook, but I’m not sure if it would actually cause a problem. With its lower resistance, 2/3 of the power would come through the cab

2. I disconnect two of the speakers in the cab, making it a 16 Ohm cab. There’s plenty of power-rating headroom available in the two other speakers. Now the amp can see 8 Ohm combined, which is OK. But do the two disconnected speakers spoil the bass by passively letting sound out by being pushed out when the active speakers are pulling in?

3. As 2, but I also add a shorting link to each of the disconnected speakers, across their terminals. This might damp their movement, making them less susceptible to being passively driven out-of-phase.

4. Physically remove two speakers and blank them off with panels of MDF or ply. Cab is now a 16 Ohm 2x12, but can be put back to original if needed.

5. Keep all the speakers active, but feed the input to the cab through an extra 8 Ohm resistor. Impedance is now 16 Ohms for the cab, so the amp is happy with both running, set to 8 Ohm output. The cab takes 20W, of which the resistor takes half. I’d make the resistor out of say 4 x 5W 33 Ohm in parallel. With this, there is some loss of output due to the resistor dissipation, which is OK because I don’t drive the amp to maximum. The resistors might affect the tone of the cab, being a non-inductive part of the load. Most of the sound output (2/3) now comes through the combo.

6. A version of 5, but with a lower resistance (hence moving towards option 1), so that both cab and combo make the same sound output. I figure the value of this resistor is 3.3 Ohms (I won’t list the theory). The amp now sees a combined resistance of 16 Ohms in parallel with 11.3 Ohms, which is 6.6 Ohms – not too far away from 8 Ohms.

7. Alternatively, with 5 Ohms in series with the cab, I can get very close to matching an 8 Ohm overall resistance (about 7.2 Ohms), and also very close to matching output between cab and combo(about a 1db difference)

8. Any other ideas?

Cheers

John

[Runewalker]

"1. I just plug the 8 Ohm cab in, in addition to the 16 Ohm internal speaker. The amp would be set to 8 Ohms but it would see 5.3 Ohms, based on the nominal resistances. This is not recommended in the handbook, but I’m not sure if it would actually cause a problem. With its lower resistance, 2/3 of the power would come through the cab
"

JH:

Not that this specifically applies to your JMarshall, but my Super Six, a Fender Twin Brain with 6 10"ers is arrayed with the 2 banks of 3 combined in Parallel and Series to create a load of 5.3 Ohms. Fender did nothing to the am to accomodate the high or low ohmage (depending if the target was 4 or 8 ohms), and it handles the load fine.

[RJB]

If you're of a mind to tinker in the cabinet, you could wire all the speakers in series (32Ù) and run that in parallel with the 16Ù internal. Net result 10Ù! Should work fine at the 8Ù setting.

As far as running at the 5Ù setting, I would believe that a newer amp with everything designed to the penny, might have problems running at a lower impedence. Mostly stress to the power tubes and final transformer. You could always look at the transformer, there could be an unconnected 4Ù tap.

Just my thoughts

[UnklMickey]

hi Ron,

putting the Celestions all in series for 32Ω , then paralleling that with 16Ω internal speaker is okay for getting the load close.

so in that respect, you had a really good idea.

but the bigger problem is distribution of power.

there will be 2/3 of the power going to the 16Ω internal speaker.

1/3 of the power will be delivered to the external cabinet, divided into the 4 speakers.

so each of the speakers in the external cabinet will see 1/8 of the power that the internal speaker sees.



i wouldn't be too concerned about the amp seeing a 5.33Ω load instead of 8.

this is a tube amp.

unlike transistor amps, tube amps are very tolerant of lower load impedances.

in fact Fender shorts the output if no speakers are connected.

no load is more dangerous to tube amp than a short.



let's look at the power distribution with John's scenario #1:

1/3 of the power goes to the internal speaker.

2/3 of the power goes to the external cabinet.

each of the speakers in the external cabinet see 1/2 as much power as the internal speaker.

not perfect, but pretty good.


putting an 8 ohm resistor in series with the external speaker is IMHO a bad idea.

it gets the impedance right on, but:

1 -- it mucks up the power distribution even worse than not having it.

2 -- it wastes power (1/4 of the total)

3 -- such a large resistance destroys the damping factor seen by the external cabinet.


if we put the internal speaker in parallel with a 2.5Ω resistor, that will look like ~2Ω.

if we put that in series with the external cabinet, we will now have 10Ω (actually 2.16Ω)

but in terms of power distribution, it falls apart.

4/5 of the power will be going to the cabinet with the 4 speakers. (good!)

but of the other 1/5, we waste about 85%. about 15% gets to the internal speaker.

if you'll pardon the expression, this is sorta percented up.



it looks to me like with these values, i'd go with scenario #1.

if the efficiency of all the speakers are equal (we don't know this!),

then we will have the external cabinet receiving twice the power,

so it should be about 3dB louder.

but i've taken way too much liberty here in making this assumption.

we don't know the relative efficiency of the internal speaker vs the ones in the cabinet.

and we haven't taken the acoustic environments into account.



unk

[RJB]

I tend to agree that option 1 is probably the best for tonal reasons. I never really did much work with tubes, so My only concerns were the stressing of components. However this is a Marshall, so one MIGHT assume they designed in some abuse factor. IIRC the '401 is the mid size combo, but I doubt they use a common transformer for any of the model, and I image it's custom for them. Not a Hammond you could look up the specs on.

If I were designing I'd design to tolerate 4Ù, but even at 5.3Ù that IS a 50% increase over the stated max. If it were me. I'd probably give it a try. Of course the engineer in me would also have the unit apart monitoring voltages and such when I did it.

BTW I bet it will sound great!
Ron

[UnklMickey]

hey Ron,

FWIW, if you are tired of having your Ω being displayed as Ù,

try this instead: Ω

when you insert that, it comes out as Ω when you preview or post.



John,

if you really like the sound of the DieSeL's speaker, and want more of it in the sound picture, there is another way.

but it would mean another trip to town in the horse-cart and some more $.

you could build a cabinet, that would provide an identical acoustic environment for another 16Ω speaker just like the one in the DieSeL.

with a special cable that would parallel those 2 speakers, and put them in series with your 4-12 cab.

that would give you exactly 8Ω

half the power would be divided between the 2 16Ω speakers and the other half would be divided between the 4 speakers of the 4-12 cab.



of course that might make for a tall stack.

and it requires another speaker.

and it requires the work to assemble another enclosure.

so this might be a bit over the top.



unk

[JohnH]

Thanks for the input guys. It seems that there is a consensus that putting the 8 Ohm cab and 16 ohm internal in parallel is not too bad, so I might cautiously try it. I’m also quite keen to try option 7, where I put a smaller resistor (5 Ohms) in series with the cab, which gets very close to achieving impedance matching and even power distribution, especially since I was quite pleased with myself for working out the math for it. At max output, the resistor takes about 8W out of the 40W. It should be low risk to try it, and then I can hear if the changes to damping are detrimental.

Alternatively, rather than build another cab, I would just do option 4 and take out and put away two of the cab drivers, to achieve a 16 Ohm cab.

John

[Runewalker]

johnh said:

Alternatively, rather than build another cab, I would just do option 4 and take out and put away two of the cab drivers, to achieve a 16 Ohm cab.
John

OK, John, now you know that is just wrong. The call a 4 driver Marshall cab a HALF-stack because it is is missing .... its other HALF. So now you want to turn it into a Quarter-Stack? James Marshall would be very disappointed in your denial of your heritage.

[JohnH]

We'll I tried option 1, just plugging the cab in along with the internal speaker. It seems to work fine, and it is a great sound, especially with the two boxes a few feet apart. A simple chugging riff satisfies most of my AC/DC and Deep Purple fantasies!

John

[JohnH]

Just to conclude on this. I got a bunch of high-wattage resistors, to try the resistor-in-series-with-the-cab options. 8 Ohms would have given me impedance matching, but audibly too much power to the combo speaker. I tried with 5 Ohms - still too much towards the combo. But with the resistance set to the theoretical balanced value of 3.3 Ohms in series with the cab, the balance between cab and combo is just right - both give the same output and it sounds great.

So I'm logging this factoid:

To get matched outputs from a 16 and an 8 Ohm speaker in parallel, put a 3.3 Ohm resistor in series with the 8 ohm.

The theoretical value is
8 x (sqrt2 - 1) = 3.3

Each speaker/cab produces 41% of the output, and the resistor dissipates 18%, or about 7W in my case.

John

[UnklMickey]

i would have guessed that 3.3Ω would have decreased the damping factor enough to make the lows to the 4-12 weak and flabby.

but...


success is.

[UnklMickey]

johnh said:

...It seems to work fine, and it is a great sound, especially with the two boxes a few feet apart....

if you're in the mood for some no-cost sonic experiments, try turning one of the boxes around, so that it faces the wall.

then move them closer and farther away.

[JohnH]

I have been finding that if the two boxes are stacked - the best sound is with some missalignment in their orientation, say 10 degrees. It makes for an untidy stack, but reduces the beam of in-yer-face treble when one is in front of it and everything is perfectly aligned.

The bass is deep and tight. I'm thinking that my extra 3.3Ohms is only a small increase in total resistance of OT windings, speakers and resistor, so no apparent extra muddiness at the low end

John