Frequency Response Measurement Question

I am running frequency response measurements on my system using a RS Digital meter and Rives Audio Test Tone CD. I use both the 1/3 octave tones and the 1/6 octave for bass response.

My question is this: I just changed amplifiers and am finding very different response curves even though:
1. Speaker placement is identical.
2. Measurement position is identical.
3. Source and cabling is identical Including power cord. (I run a dedicated circuit and wall outlet for the amp and DAC. Everything else is on a separate circuit)
4. Power rating of amps is very similar.

Why is this given both amps are fairly high quality and should produce a fairly flat frequency response? Since the setup is nearly identical, it seems this is not the room response but rather a difference in the amplifiers.

The problem is, I have a fairly flat response but with a big bump at 2000 hz - which is a noticeably bad place to have a loud spot, in other ways high frequency is flatter with the new amp. Any ideas on how to address this.

Green Mountain Europa Speakers
Monarchy DAC 24 (Solid State section used for both)
Integra CD player for transport.
Paradigm PW-2100 sub, converted to down firing.
Cobalt Ultimate Speaker cables
Blue Jeans interconnects

1st amp - Arcam AVR 300 in direct mode, 2 channels used, 120 watts approx..
I actually really liked the sound of this amp but wanted to simplify since I no longer am building home theater.

New amp - Simaudio I-5, 110 watts at 4 ohms.
In this setup, this amp has more clarity and sounds more powerful (dynamic?) but less transparency with less separation between instruments and exhibits a harshness (the 2k bump I suspect). So if I could find a way to fix that, I think it would sound great.

Thanks in advance for your input!!!
I'd suggest to check your bass management - perhaps settings were not identical as you first thought. As for high frequencies, these are quite location dependent - is it possible you moved the meter a few inches between measurements?
As Shadorne says, moving the meter even a fraction of an inch can cause large changes in higher frequencies especially when comb filtering is present. Just move you head side to side over the meter at 2000 hz and if it is present, you will hear it easily.

Even if you didn't move the measurement position at all, the location of your body can change the measurements dramatically at higher frequencies. As you move around the meter, you shield the meter from some direct reflections and/or cause new reflections towards the meter. Try it and see.

Does the Rives disk use warble tones? These help to minimize these effects.

I have come to use the RS meter only under 200 hz and use a Behringer for viewing the entire FFT spread spectrum. The ramdon frequency pink noise tends to "average" over the response and allow you to see how flat it is without seeing all of the upper frequency peaks and valleys which will always be there.
Good comments - It is possible that things moved inches. I also have noticed the effect of my body position on the high frequencies. I will go back and try some different positions to see if the 2k problem is placement.

I spent all day today working on the bass frequencies and actually found a big problem with my settings that I have corrected. I do believe the sound is improving bit by bit.