Sherwood RX-4109 Measurements

Pete B

AK Member
Subscriber
I've been interested in this amp for some time for basic high powered systems but was concerned about driving 4 ohm loads. Posted some comments here about the design and output devices used:
http://www.audiokarma.org/forums/showpost.php?p=8748108&postcount=71

http://www.audiokarma.org/forums/showpost.php?p=8748587&postcount=72

And:
http://www.audiokarma.org/forums/showpost.php?p=8751045&postcount=73

Decided to measure the unit, all voltages measured with:
Fluke 87 true RMS multimeter
HP function generator for sine source
Cheap B&K scope for viewing clipping
Loaded both channels with 8 ohm 1% 100W resistors made up of a string of four 2 ohm resistors, 25W each.
Did not set up to measure distortion, it is generally recognized that a good looking sine by eye on a scope should be less than 1% distortion, and just visible clipping is often also just below 1% distortion.

Set it up for testing and first measured the power line voltage:
idle = 120.9 Vac
both channels driven 100W/ch = 118.8
Line voltage is steady enough that a Variac is not needed.

Measured the raw DC at the 6800 uF 65V power caps at:
B+ = 57.8
B- = 57.7

Measured RMS output voltage and calculated power are given in the attached image.
It is as expected, just meets spec in the middle of the band, and is perhaps a bit low at 20Hz.
Checked Sherwood's spec and it is for 105 W 40 to 20 KHz less than .9 percent THD, and I'd say
that it meets the spec since the onset of clipping measurements are all over 105W and the rating
is for .9% distortion.
The reduced output at 20 Hz is due to the small (for the rated power) PSU caps at 6800 uF each.
The caps are not able to maintain the supply voltage between peaks of the transformer rectified output,
this is completely normal and larger caps would help. 10,000 uF each would be a decent upgrade
20,000 uF if 4 ohm loads are going to be used - not recommended.


I did not want to overly stress and break this unit since it is obviously a budget model not intended for extended full power use. I brought up the power just long enough to take the measurement, then backed it down. Even so the heat sink got VERY HOT, I could hold my hand on it but I really did not want to leave it there. I should get a temperature probe. No matter what these kinds of temps are not good, it clearly needs either a bigger heat sink or fans.

Interesting to note that this unit has an extra thermal sensor on the heat sink that I assume is to shut it down to prevent output device failures. Does anyone know if the older ones had the sensor? It is toward the middle of the heat sink and has a black and white wire leading from it. If this is a new feature that prevents failures when driving 4 ohm loads or when an object is accidentally left on top of the unit blocking cooling then one does not have to be as concerned about reports of failed outputs.
 

Attachments

  • SHERWOOD-POWER-TEST.png
    SHERWOOD-POWER-TEST.png
    14 KB · Views: 25
Last edited:
Attached are pictures showing:
Top view
Output Transistors and Thermal Sensor
Power Transformer
PSU Caps
Old Dell Tower Heat Sink, a pair might fit bolted on right above the output devices. I have to buy a pair and try it. There is a bit over 3" from the top of the transistors to the cover.

Capacitors are Samyoung brand, never heard of them and they don't give me a lot of confidence.
At least the unit should be easy to recap!
 

Attachments

  • Dell-TOWER-HEAT-SINK.JPG
    Dell-TOWER-HEAT-SINK.JPG
    66.3 KB · Views: 39
  • 100_0830.jpg
    100_0830.jpg
    113 KB · Views: 48
  • 100_0829.jpg
    100_0829.jpg
    108.2 KB · Views: 45
  • 100_0828.jpg
    100_0828.jpg
    107.9 KB · Views: 51
  • 100_0827.jpg
    100_0827.jpg
    113.5 KB · Views: 55
Last edited:
Dynamic Power

Previously stated:
Measured the raw DC at the 6800 uF 65V power caps at:
B+ = 57.8
B- = 57.7

If we assume that the amp does not current limit, either due to protection (it does not have this type) or due to beta droop in the output devices then we can compute the Dynamic Power as follows, this would be the approximate power if the PSU was regulated at the idle voltage:
The peak waveform output voltage is a few volts less than the supply due to Vce sat of the outputs and drivers:
Let's say Vpeak = 54 volts
then Vrms = .707 * Vpeak or 38.2 Vrms
Power = V^2/R = 38.2^2/8 = 182.4 Watts
Expressed as dB over the 105 rated power is then 10 * log (182.4/105) = 2.4 dB
This is excellent and larger caps will help sustain this peak power for a longer period of time.
This is not yet confirmed, have to get my burst generator going to see if it in fact does not current limit.
 
Here is a picture with a left over $3 computer case fan set on top of the heat sink, there is plenty of room. Rather than put in a speed control, I'd just put a pair of 12V fans in series and run the pair at half voltage from a 12 or 15 V supply. Any amount of forced air movement pushes heat sink efficiency way up.

Second pic shows some 15,000 uF 105 deg C caps that I had on hand. They'd make a nice upgrade.
 

Attachments

  • 100_0832.jpg
    100_0832.jpg
    104.8 KB · Views: 44
  • 100_0833.jpg
    100_0833.jpg
    117.3 KB · Views: 44
Last edited:
One more observation that I forgot to mention. I started out with the volume at about 30 and found that the output clipped far below full power, obviously a preamp section before the volume control was clipping. When I turned the volume above 60 everything was fine.
Keep this in mind if you plan to use it more as a power amp, keep the volume high or even on max if you say drive a line input from another preamp or device with its own volume control.
 
Last edited:
You can see here on William's Teardown page that his version from around 2010 does not have the additional thermal sensor that I mentioned above:
http://greyghost.mooo.com/rx4109takeapart/

Also, the layout is somewhat different.

Looking side by side, the phono amp is no longer a daughter card off of the main board, and the tuner can is much smaller, hope it performs as well.



I'm tempted to see what it takes to trip that thermal sensor!

Also, if I was being paid to evaluate the design, I'd short the output at full power after it has gotten very hot.
 
Last edited:
Here is a Dell 9Y692 heat sink that I had lying around. There is about .25 inch clearance to the lid when it is about 1/8" above the top of the transistors. The contact area is flat so there should be fairly good conduction from the main heat sink. I'd rather not have to remove the 4109 heat sink to drill it, not sure of a good way to mount it.

This Thermal Adhesive gets excellent reviews, but I worry about the inertia of all that weight, if falling off and shorting something inside the case:
http://www.arcticsilver.com/arctic_alumina_thermal_adhesive.htm

Removing and drilling for screws is probably the way to go.

Working on the power transistors would then become very difficult without removing the additional heat sink.
 

Attachments

  • 100_0834.jpg
    100_0834.jpg
    121.4 KB · Views: 38
The above processor heat sink is probably too large, also the tight fin pitch is designed for fan cooling, they are usually wider for convection cooling.
I found what I think are Northbridge heat sinks on ebay and it seems that 4 of these might fit nicely and being less tall should allow good enough access to the output transistor screws. I might buy a few of them and give it a try. Also, being lighter gluing them on should work.

I'd prefer to stay with passive cooling but fans are looking better if there is no good way to mount helper heat sinks.

Any ideas for mounting heatsinks or a good simple fan setup?
 

Attachments

  • SMALL-HEATSINK1.JPG
    SMALL-HEATSINK1.JPG
    36.7 KB · Views: 11
  • SMALL-HEATSINK2.JPG
    SMALL-HEATSINK2.JPG
    55.1 KB · Views: 11
Last edited:
Found these 2SB1647 and 2SD2560 as slightly higher power and current (130W/15A) versions of the output devices (100W/10A):
http://www.semicon.sanken-ele.co.jp/sk_content/2sd2560_ds_en.pdf

Considering the rail voltages these should always be used for repairs but even better would be 200W devices if there were Darlingtons. I'd expect these to be drop in being Darlingtons and having about the same Ft.

I'm interested in taking a look at a 4109 with blown outputs if anyone has one in the Connecticut area, I'll buy it for a reasonable price.
 
Last edited:
I'm thinking of this receiver for a few friends who want to get into 2 ch sound at minimal cost. Thinking of pairing this with Large Advents, but they are about 6 ohms with the NLA dipping below even 3 ohms in the tweeter range - not good. I can't suggest this amp with NLAs unless they are used in the decrease position only where more resistance is put in the tweeter path.
 
I am pretty sure this receiver can drive NLAs with no problem as long as you have not modded anything else besides adding a heat sink. Of course just one pair of speakers.
I have the RX-5502.
 
I've measured the NLA input impedance and posted some of the data at the Classic Speaker Pages. The RX-4109 is not going to like the dip below 3 ohms in the Extended position. I've not measured the Sherwood with such low Z loads so I don't know if the power is reduced, it distorts badly, current limits, or if it will burn out and fail, or any combination.
We can't know for sure until someone does some measurements on the Sherwood.
 
If I am not mistaken Uxwbill has driven 2 sets of 8 ohm speakers with these. Not recommended though. NLAs are rated at 8 ohms and the VC resistance should measure around 5 ohms.
I think you are being overly cautious.

I was at Radio Shack today and they are on sale for $74.99.

I fried a cheapo Teac receiver once but it was because the woofer on my Advent/3 went south.

Live dangerously...hook up the NLAs and see what happens.
 
I'm going to test the amp first, and not doing any experiments with my Advent collection until I know what the amp can do. He used a pair of 8 ohm speakers for a combined value of 4 ohms and must not have driven it hard since he said it didn't get hot.

Doesn't matter how NLA's are rated, the fact is that they are nominally 6 ohms but dip
below 3 ohms in the extended position, 2nd time I've said this.

The RX-4109 was discounted to $50 at the lowest point.
 
NLAs are about 91db....should be pretty easy on the Sherwood. Maybe I will hook up my Advent/1s to my 5502 and see what happens. I usually treat my 6 ohm speakers the same as 8 ohm. Of course I would not run four 6 ohm speakers at the same time.
 
LAs and NLAs are not 91 dB 2.83 V sensitivity or efficiency, see efficiency here at 83 dB, 2.83 V is about 2 dB more at about 85 - 86:
http://www.enjoythemusic.com/magazine/bas/0808/

Or here for the Small Advent, 2.83V sensitivity comes in at about 85 LA/NLA are about 2 dB better giving 87 dB:
http://www.stereophile.com/content/smaller-advent-loudspeaker-measurements

Note that the Small Advent is rated at 4 ohm and never dips below that value.
LA is about 6 ohms.
NLA is also about 6 ohms in the woofer range but dips below 3, on Extended, in the tweeter range.
 
Last edited:
My last 4109 is working in my basement powering my Polk Monitor 4 speakers. My son-in-law has my other 4109. I did lose one 4109 years ago because I didn't follow the clearences recommended in the manual. I still think it is a great buy for performance vs price.
 
I just bought an Onkyo TX 8522 for my son at a very good price.
It is a similar 100W ch, 2ch receiver but is rated for 4 ohm loads and the dynamic
power is rated into 2 ohm loads. Very nice.
 
Back
Top Bottom