Did you know that not all protectors are created equal? Had one die. Telco guy went to spare, but didn't remove or mark the old one. Not to happy about that.

The symptom was Noise for dial tone and about 5 VAC across the line terminals when off hook. Didn't occur to me it could be a protector.

I changed it to a Bourns model and my DSL speeds went up to about what it should be. Nearly doubled.

Although the 50 YO vintage suppressor is still in place and technically it isn't needed, but I still haven't had a chance to remove it and see if speeds could improve further.

What do you think?

KISS, not sure if this would be of any assistance due to differences, but here's how we'd tweak the wiring for UK DSL connections: DSL ZoneUK - Improve your speed - Extension wiring and Disconnection help

I'm tweeked as best as I can be for 3MB/s/ 768kbps.

Splitter vs filters
No 1/2 ringer in NID
Jumperless test jack in NID
Modem 5' from NID (Bridge mode)

Negatives:
2 protectors (Can be good or bad)
All other wiring QUAD wiring with a mix of home run and bus (Yuk)

Router in center of house.
Access Point in center of house
One wireless repeater

I just happened to notice a degradation in speed until it finally failed.

There is even a protector with matched capacitance available for ADSL2.

So, what I am saying based on real and manufacturer's data is that the protector (surge suppressor) can impact speed.

I think the Carbon protectors would be more sucepptible to noise than gas.
I use the Bournes, and prefer a whole house filter.( This eleminates the need for filters everywhere).
Going to Curleybens link now, Take Care

I used to design this stuff for Bell Labs...

Modern gas tube protectors are definitely much, much better at passing high-speed signals with minimal interference compared to the old carbon blocks. It wasn't until the early 80's that any of the people designing these things ever worried about carrying anything above about 4K hertz. The old carbon blocks when compared to modern plug-ins are much more susceptible to noise - especially if you have the really old screw type, instead of the 5-pin or 3-pin style - because of shorts to ground through old & dirty carbon blocks as well as poor wiring practices that separate tip from ring, which can cause crosstalk from adjacent lines or other external sources. These old bocks are typically wired with tip and ring untwisted, which is a definite no-no for modern high-speed service like DSL. I once did an experiment that showed that about two feet of untwisted wiring through a building entrance protector caused more crosstalk between pairs for DSL or ISDN lines than you get through 5000 feet of OSP cable. In DSL applications this results in slower through speeds.

And yes, subscriber protectors do indeed die. After a few high voltage spikes the old carbon block style can start sparking to ground, which would appear to the consumer as a noisy line. Gas tubes typically last much longer, but they can leak, and so in most cases have a back up gap that is supposed to provide a fail safe. When these protctors die they typically do it quietly - leaving your line poorly protected but with no noise or loss of service to alert you.

Is it agreed that temporary removal of protectors to verify a defective protector?

I'd say the answer is NO, but only because it will show you have a bad one, but not a good one, In the ideal case, it should shorted, but as ebaines pointed out glass leaks air and there for the discharge charachteristics can become air rather than the gas mixture and this gives a false sense of security.

The only real way to test, would be to test breakdown with say the specific Bourns tester. Leakage should show on a DVM, but measuring 10 G ohm is not within the realm of most people. Capacitance can also be checked and you could reject based on that.

1st order test - temporarily disconnect

Ebanes:

What I actually have is a Bourns 455 protector outside at the NID and a splitter at the NID. Then the telco line goes to an old carbon protector/junction block. Basically the protector there isn't really doing much good, but really should not be impacting the DSL signal.

It's the old style carbon protector/junction block with terminals about 3" apart. All quiad wire from then on.

So, the new ones are basically two (gas discharge tube in parallel with a ZNR) to ground and possibly one for comon mode signals? The gas discharge tube handles the fast strikes and the ZNR SHOULD fail shorted, but doesn't always.

I'm not familiar with that Bourns protector, but I can tell you that over the past ten years or so manufacturers have moved away from using two separate gas tubes one for tip-to-ground and the other for ring-to-ground) and in favor of a single tube. The advantage of the newer design is that if one lead gets a high enough voltage the tube fires and both leads are automatically grounded, whereas with separate tubes you can have a situation where only oe tube fires, which can cause a high current surge from the other lead through the customer premises equipment to ground.

I bet if you removed the old protector block and neatened the wires you just might see an improvement in transmission speed.